-
PlantScience.ai: An LLM-Powered Virtual Scientist for Plant Science
Haopeng Yu, Shasha Zhou, Mingyu Huang, Ling Ding, Yuxuan Chen, Yinru Wang, Yingyu Ren, Nuo Cheng, Xinya Wang, Jie Liang, The John Innes Centre and The Sainsbury Laboratory Collaboration, Huakun Zhang, Yiliang Ding, Ke Li
Molecular Plant (MP)
10.1016/j.molp.2026.03.010
Abs | PDF | Web | BiB | Cited by 0
The accelerating growth of plant science knowledge presents a major challenge for researchers seeking to extract accurate, up-to-date knowledge from an increasingly fragmented and domain-specific corpus. General-purpose large language models (LLMs), while powerful, often misinterpret plant science terminology and lack mechanisms for source traceability. We created PlantScience.ai, a virtual plant biology scientist powered by our automated scientific knowledge graph construction pipeline (AutoSKG). PlantScience.ai exhibits expert-level reasoning in plant biology and maintains scholarly rigour in its citations. Through continuous learning, it integrates the latest research, ensuring that its knowledge base remains current and scientifically robust. Apart from providing the answers to the scientific questions, PlantScience.ai can interact with human scientists, follow instructions, and retrieve information with citation awareness, grounding each response in primary sources to ensure accuracy and verifiability. PlantScience.ai marks a pivotal advance toward a collaborative scientific paradigm in which virtual and human plant scientists work synergistically to accelerate discovery while preserving the unique value of human insight. PlantScience.ai is available at https://plantscience.ai.
@article{Yu2026,
author = {Haopeng Yu and
Shasha Zhou and
Mingyu Huang and
Ling Ding and
Yuxuan Chen and
Yinru Wang and
Yingyu Ren and
Nuo Cheng and
Xinya Wang and
Jie Liang and
Huakun Zhang and
Yiliang Ding and
Ke Li},
title = {PlantScience.ai: An LLM-Powered Virtual Scientist for Plant Science},
journal = {Molecular Plant},
year = {2026},
url = {https://www.cell.com/molecular-plant/fulltext/S1674-2052(26)00080-8},
doi = {10.1016/j.molp.2026.03.010}
}
-
Adaptive Population Classification based Multi-Strategy Evolutionary Algorithm for Dynamic Constrained Multi-Objective Optimization
Xueqing Wang, Biao Luo, Zhanglu Hou, Juan Zou, Jinhua Zheng, Ke Li
Expert Systems with Applications (ESWA)
10.1016/j.ijepes.2026.111774
Abs | BiB | Cited by 0
Dynamic constrained multi-objective optimization problems (DCMOPs) commonly arise in practical scenarios, where the time-varying nature of constraints introduces significant challenges. A major limitation of most existing methods is their over-reliance on historical feasible solutions while overlooking the guidance that can be derived from infeasible ones. This may lead to rapid diversity loss and poor adaptability, especially when environmental changes result in discontinuous feasible regions. To overcome this limitation, this paper proposes an adaptive population classification-based multi-strategy evolutionary algorithm (ATCDA) to solve DCMOPs. Specifically, ATCDA first classifies the historical population into three distinct subpopulations based on constraint violation and crowding distance. It then designs three specialized strategies to regenerate each subpopulation for the new environment. The feasible subpopulation is optimized via autoencoder-based prediction; the high-crowding ranking infeasible subpopulation undergoes objective correction; and an adaptive Gaussian mutation is designed to the remaining individuals. Finally, these three newly regenerated subpopulations are merged to form an initial population, enabling an effective response to the change. Experimental results on a set of DCMOPs show that the proposed ATCDA is competitive compared to other state-of-the-art methods.
@article{WangLHZZL26,
author = {Jiangjiao Xu and
Long Ma and
Ke Li and
Dongdong Li and
Leijiao Ge and
Changjun Jiang},
title = {Semantic-probabilistic co-optimization framework for distributed non-linear optimal power flow},
journal = {International Journal of Electrical Power & Energy Systems},
year = {2026},
volume = {177},
pages = {111774},
url = {https://www.sciencedirect.com/science/article/pii/S0142061526002164},
doi = {10.1016/j.ijepes.2026.111774}
}
-
Semantic-Probabilistic Co-Optimization Framework for Distributed Non-linear Optimal Power Flow
Jiangjiao Xu, Long Ma, Ke Li, Dongdong Li, Leijiao Ge, Changjun Jiang
International Journal of Electrical Power and Energy Systems (IJEPES)
10.1016/j.ijepes.2026.111774
Abs | BiB | Cited by 0
With the increasing integration of renewable energy sources, power systems are becoming more dynamic and decentralized. However, traditional distributed Optimal Power Flow (OPF) methods often rely on linearized approximations, limiting their ability to handle the nonlinear nature of modern power networks. To address this, a distributed Optimal Power Flow (DisOPF) approach based on the Alternating Direction Method of Multipliers (ADMM) is proposed, enhanced with Bayesian Optimization (BO) to handle nonlinearity and enable adaptive parameter tuning. By integrating the Bayesian Optimization model into the ADMM update process, the approach effectively avoids the approximation errors in node modeling typically introduced by ADMM. Additionally, the BO model reduces the repeated computation of losses during ADMM iterations, thereby accelerating convergence and improving the solution quality. In addition, Large Language Models (LLMs) are employed to analyze historical data and assist in parameter selection through context-aware suggestions. Simulation results on the IEEE 33-nodes and IEEE 9-nodes systems demonstrate that the proposed method enhances convergence speed and voltage coordination, while maintaining robustness under diverse initializations and load disturbances. These results highlight the potential of combining probabilistic optimization and language models for distributed power system optimization.
@article{XuMLLGJ26,
author = {Jiangjiao Xu and
Long Ma and
Ke Li and
Dongdong Li and
Leijiao Ge and
Changjun Jiang},
title = {Semantic-probabilistic co-optimization framework for distributed non-linear optimal power flow},
journal = {International Journal of Electrical Power & Energy Systems},
year = {2026},
volume = {177},
pages = {111774},
url = {https://www.sciencedirect.com/science/article/pii/S0142061526002164},
doi = {10.1016/j.ijepes.2026.111774}
}
-
RIDER: 3D RNA Inverse Design with Reinforcement Learning-Guided Diffusion
Tianmeng Hu, Yongzheng Cui, Biao Luo, Ke Li
Proc. of the 14th International Conference on Learning Representations (ICLR'26)
PDF | BiB | ≈ 28%
-
Augmenting Biological Fitness Prediction Benchmarks with Landscapes Features from GraphFLA
Mingyu Huang, Shasha Zhou, Ke Li
Proc. of the 39th Annual Conference on Neural Information Processing Systems (NeurIPS'25), Dataset and Benchmark Track
Spotlight paper (Top 2.8%)
PDF | Blog | BiB | ≈ 24.91%
-
Multi-Fidelity Methods for Optimization: A Survey
Ke Li, Fan Li
ACM Computing Surveys (CSUR)
10.1145/3801959
Abs | PDF | Supp | BiB | Cited by 0
Real-world black-box optimization often involves time-consuming or costly experiments and simulations. Multi-fidelity optimization (MFO) stands out as a cost-effective strategy that balances high-fidelity accuracy with computational efficiency through a hierarchical fidelity approach. This survey presents a systematic exploration of MFO, underpinned by a novel text mining framework based on a pre-trained language model. We delve deep into the foundational principles and methodologies of MFO, focusing on three core components—multi-fidelity surrogate models, fidelity management strategies, and optimization techniques. Additionally, this survey highlights the diverse applications of MFO across several key domains, including machine learning, engineering design optimization, and scientific discovery, showcasing the adaptability and effectiveness of MFO in tackling complex computational challenges. Furthermore, we also envision several emerging challenges and prospects in the MFO landscape, spanning scalability, the composition of lower fidelities, and the integration of human-in-the-loop approaches at the algorithmic level. We also address critical issues related to benchmarking and the advancement of open science within the MFO community. Overall, this survey aims to catalyze further research and foster collaborations in MFO, setting the stage for future innovations and breakthroughs in the field.
@article{LiL26,
author = {Ke Li and
Fan Li},
title = {Multi-Fidelity Methods for Optimization: A Survey},
journal = {ACM Computing Surveys},
year = {2026},
url = {https://dl.acm.org/doi/10.1145/3801959},
doi = {10.1145/3801959}
}
-
Meta Representation-based Personalized Federated Continual Learning in Edge Computing Systems
Siwei Zheng, Jia Hu, Geyong Min, Ke Li
IEEE Trans. Computational Social Systems (TCSS)
10.1109/TCSS.2025.3582364
Abs | BiB | Cited by 0
Federated learning (FL) enables multiple clients to collaboratively train a shared global model while keeping their data local, fostering privacy-preserving innovation in sectors such as healthcare, finance, and telecommunications. Federated continual learning (FCL) extends FL by enabling clients to incrementally learn and adapt the global model over time, addressing dynamic data distributions and evolving tasks while maintaining data privacy. However, as differences in data distribution and class diversity across clients’ data streams widen, the local personalization and continuous adaptation of the global model to both previously learned tasks and current tasks in clients’ local data streams present a significant challenge. To address this challenge, we propose a meta representation based class incremental learning (Meta-RBCIL) algorithm for peronalized FCL in edge computing systems. The proposed approach meta-trains a representation network using FL, equipping it with strong generalization capabilities across diverse tasks. This network is then leveraged to adapt to each client’s current task data and exemplar memory using balanced gradients. By doing so, the method mitigates catastrophic forgetting in clients’ continual learning models and alleviates bias caused by the imbalance between old samples in exemplar memories and new class samples in current tasks. The experimental results show that our approach surpasses key baseline algorithms in terms of average task accuracy per communication round, across varying levels of task similarity in clients’ data streams. Furthermore, the personalized continual learning models produced by the Meta-RBCIL algorithm achieve the lowest average task forgetting ratio and the highest average task generalization capability, making Meta-RBCIL highly suitable for deployment in edge computing systems.
@article{ZhengHML25,
author = {Siwei Zheng and
Jia Hu and
Geyong Min and
Ke Li},
title = {Meta Representation-Based Personalized Federated Continual Learning in Edge Computing Systems},
journal = {IEEE Transactions on Computational Social Systems},
year = {2025},
pages = {1--12},
url = {https://ieeexplore.ieee.org/document/11081891},
doi = {10.1109/TCSS.2025.3582364}
}
-
A Survey of Multi-objective Evolutionary Algorithm Based on Decomposition: Past and Future
Ke Li
IEEE Trans. Evolutionary Computation (TEVC)
10.1109/TEVC.2024.3496507
Abs | PDF | Supp | BiB | Cited by 30
Decomposition has been the mainstream approach in the classic mathematical programming for multi-objective optimization and multi-criterion decision-making. However, it was not properly studied in the context of evolutionary multi-objective optimization until the development of multi-objective evolutionary algorithm based on decomposition (MOEA/D). In this article, we present a comprehensive survey of the development of MOEA/D from its origin to the current state-of-the-art. In order to be self-contained, we start with a step-by-step tutorial that aims to help a novice quickly get onto the working mechanism of MOEA/D. Then, selected major developments of MOEA/D are reviewed according to its core design components including subproblem formulations, selection mechanisms and reproduction operators. Besides, we also overviews some further developments for constraint handling, large-scale problems, computationally expensive objective functions, preference incorporation, and real-world applications. In the final part, we shed some lights on emerging directions for future developments.
@article{Li24,
author = {Ke Li},
title = {A Survey of Multi-objective Evolutionary Algorithm Based on Decomposition: Past and Future},
journal = {IEEE Transactions on Evolutionary Computation},
year = {2024},
pages = {1--1},
url = {https://ieeexplore.ieee.org/document/10750458},
doi = {10.1109/TEVC.2024.3496507}
}
-
DaNuoYi: Evolutionary Multi-Task Injection Testing on Web Application Firewalls
Ke Li, Heng Yang, Willem Visser
IEEE Trans. Software Engineering (TSE), 51(9): 2412–2431, 2025
10.1109/TSE.2023.3343716
Abs | PDF | Code | BiB | Cited by 6
Web application firewall (WAF) plays an integral role nowadays to protect web applications from various malicious injection attacks such as SQL injection, XML injection, and PHP injection, to name a few. However, given the evolving sophistication of injection attacks and the increasing complexity of tuning a WAF, it is challenging to ensure that the WAF is free of injection vulnerabilities such that it will block all malicious injection attacks without wrongly affecting the legitimate message. Automatically testing the WAF is, therefore, a timely and essential task. In this paper, we propose D a N uo Y i , an automatic injection testing tool that simultaneously generates test inputs for multiple types of injection attacks on a WAF. Our basic idea derives from the cross-lingual translation in the natural language processing domain. In particular, test inputs for different types of injection attacks are syntactically different but may be semantically similar. Sharing semantic knowledge across multiple programming languages can thus stimulate the generation of more sophisticated test inputs and discovering injection vulnerabilities of the WAF that are otherwise difficult to find. To this end, in D a N uo Y i , we train several injection translation models by using multi-task learning that translates the test inputs between any pair of injection attacks. The model is then used by a novel multi-task evolutionary algorithm to co-evolve test inputs for different types of injection attacks facilitated by a shared mating pool and domain-specific mutation operators at each generation. We conduct experiments on three real-world open-source WAFs and six types of injection attacks, the results reveal that D a N uo Y i generates up to 3:8× and 5:78× more valid test inputs (i.e., bypassing the underlying WAF) than its state-of-the-art single-task counterparts and the context-free grammar-based injection construction.
@article{LiYV23,
author = {Ke Li and
Heng Yang and
Willem Visser},
title = {DaNuoYi: Evolutionary Multitask Injection Testing on Web Application Firewalls},
journal = {IEEE Transactions on Software Engineering},
year = {2025},
volume = {51},
number = {9},
pages = {2412--2431},
url = {https://ieeexplore.ieee.org/abstract/document/10372386},
doi = {10.1109/TSE.2023.3343716}
}
-
MBL-CPDP: A Multi-objective Bilevel Method for Cross-Project Defect Prediction
Jiaxin Chen, Jingliang Ding, Kay Chen Tan, Jiancheng Qian, Ke Li
IEEE Trans. Software Engineering (TSE), 51(8): 2305–2328, 2025
10.1109/TSE.2025.3577808
Abs | BiB | Cited by 0
Cross-project defect prediction (CPDP) leverages machine learning (ML) techniques to proactively identify software defects, especially where project-specific data is scarce. However, existing CPDP approaches suffer from three critical limitations: ineffective exploration of high-dimensional parameter spaces, poor adaptability across diverse projects with heterogeneous data distributions, and inadequate handling of feature redundancy and distribution discrepancies between source and target projects. To address these challenges, we formulate CPDP as a multi-objective bilevel optimization (MBLO) method, dubbed MBL-CPDP. Our approach comprises two nested problems: the upper-level, a multi-objective combinatorial optimization problem, enhances robustness by optimizing ML pipelines that integrate feature selection, transfer learning, and classification techniques, while the lower-level problem fine-tunes their hyperparameters. Unlike traditional methods that employ fragmented optimization strategies or single-objective approaches that introduce bias, MBL-CPDP provides a holistic, end-to-end optimization framework. Additionally, we propose an ensemble learning method to better capture cross-project distribution differences and improve generalization across diverse datasets. An MBLO algorithm is then presented to effectively solve the formulated MBLO problem. To evaluate MBL-CPDP’s performance, we compare it with five automated ML tools and 50 CPDP techniques across 20 projects. Extensive empirical results show that MBL-CPDP outperforms the comparison methods, demonstrating its superior adaptability and comprehensive performance evaluation capability.
@article{ChenDTQL25,
author = {Jiaxin Chen and
Jinliang Ding and
Kay Chen Tan and
Jiancheng Qian and
Ke Li},
title = {MBL-CPDP: A Multi-Objective Bilevel Method for Cross-Project Defect Prediction},
journal = {IEEE Transactions on Software Engineering},
year = {2025},
volume = {51},
number = {8},
pages = {2305--2328},
url = {https://ieeexplore.ieee.org/document/11029502},
doi = {10.1109/TSE.2025.3577808}
}
-
Evolutionary Alternating Direction Method of Multipliers for Constrained Multi-Objective Optimization with Unknown Constraints
Shuang Li, Ke Li, Wei Li, Ming Yang
IEEE Trans. Evolutionary Computation (TEVC), 29(4): 1419–1433, 2025
10.1109/TEVC.2024.3425629
Abs | PDF | BiB | Cited by 13
Constrained multiobjective optimization problems (CMOPs) pervade real-world applications in science, engineering, and design. Constraint violation (CV) has been a building block in designing evolutionary multiobjective optimization (EMO) algorithms for solving CMOPs. However, in certain scenarios, constraint functions might be unknown or inadequately defined, making CV unattainable and potentially misleading for the conventional constrained EMO algorithms. To address this issue, we present the first of its kind evolutionary optimization framework, inspired by the principles of the alternating direction method of multipliers that decouples objective and constraint functions. This framework tackles CMOPs with unknown constraints by reformulating the original problem into an additive form of two subproblems, each of which is allotted a dedicated evolutionary population. Notably, these two populations operate toward complementary evolutionary directions during their optimization processes. In order to minimize discrepancy, their evolutionary directions alternate, aiding the discovery of feasible solutions. Comparative experiments conducted against the five state-of-the-art constrained EMO algorithms on 120 benchmark test problem instances with varying properties as well as two real-world engineering optimization problems demonstrate the effectiveness and superiority of our proposed framework. Its salient features include faster convergence and enhanced resilience to various Pareto front shapes.
@article{LiLLY25,
author = {Shuang Li and
Ke Li and
Wei Li and
Ming Yang},
title = {Evolutionary Alternating Direction Method of Multipliers for Constrained Multiobjective Optimization With Unknown Constraints},
journal = {IEEE Transactions on Evolutionary Computation},
year = {2025},
volume = {29},
number = {4},
pages = {1419--1433},
url = {https://ieeexplore.ieee.org/document/10589587},
doi = {10.1109/TEVC.2024.3425629}
}
-
Evolutionary Art Attack For Black-Box Adversarial Example Generation
Phoenix Williams, Ke Li, G. Min
IEEE Trans. Evolutionary Computation (TEVC), 29(4): 1343–1355, 2025
10.1109/TEVC.2024.3391063
Abs | PDF | Code | BiB | Cited by 2
Deep neural networks (DNNs) have achieved remarkable performance in various tasks, including image classification. However, recent research has revealed the susceptibility of trained DNNs to subtle perturbations introduced into input images. Addressing these vulnerabilities is pivotal, leading to a significant area of study focused on developing attack algorithms capable of generating potent adversarial images. In scenarios where access to gradient information is restricted (black-box scenario), many existing methods introduce optimized perturbations to each individual pixels of an image to cause trained DNNs to mis-classify. However, due to the high-dimensional nature of this approach, current methods have inherent limitations. In contrast, our proposed approach involves the construction of perturbations by concatenating a series of overlapping semi-transparent shapes. Through the optimization of these shapes’ characteristics, we generate perturbations that result in the desired misclassification by the DNN. By conducting a series of attacks on state-of-the-art DNNs trained of CIFAR-10 and Imagenet datasets, our method consistently outperforms existing attack algorithms in terms of both query efficiency and success rate.
@article{WilliamsLM25,
author = {Phoenix Neale Williams and
Ke Li and
Geyong Min},
title = {Evolutionary Art Attack for Black-Box Adversarial Example Generation},
journal = {IEEE Transactions on Evolutionary Computation},
year = {2025},
volume = {29},
number = {4},
pages = {1343--1355},
url = {https://ieeexplore.ieee.org/document/10504779},
doi = {10.1109/TEVC.2024.3391063}
}
-
Optimal Parameter Adaptation for Safety-Critical Control via Safe Barrier Bayesian Optimization
Shengbo Wang, Ke Li, Zheng Yan, Zhenyuan Guo, Guanghui Wen, Shiping Wen
IEEE Trans. Control Systems Technology (TCST), 33(5): 1953–1959, 2025
10.1109/TCST.2025.3561059
Abs | PDF | Code | BiB | Cited by 1
Safety is of paramount importance in control systems to avoid costly risks and catastrophic damages. The control barrier function (CBF) method, a promising solution for safety-critical control, poses a new challenge of enhancing control performance due to its direct modification of original control design and the introduction of uncalibrated parameters. In this work, we shed light on the crucial role of configurable parameters in the CBF method for performance enhancement with a systematical categorization. Based on that, we propose a novel framework combining the CBF method with Bayesian optimization (BO) to optimize the safe control performance. Considering feasibility/safety-critical constraints, we develop a safe version of BO using the barrier-based interior method to efficiently search for promising feasible configurable parameters. Furthermore, we provide theoretical criteria of our framework regarding safety and optimality. An essential advantage of our framework lies in that it can work in model-agnostic environments, leaving sufficient flexibility in designing objective and constraint functions. Finally, simulations on swing-up control and high-fidelity adaptive cruise control (ACC) are conducted to demonstrate the effectiveness of our framework.
@article{WangLYGWW25,
author = {Shengbo Wang and
Ke Li and
Zheng Yan and
Zhenyuan Guo and
Song Zhu and
Guanghui Wen and
Shiping Wen},
title = {Optimal Parameter Adaptation for Safety-Critical Control via Safe Barrier Bayesian Optimization},
journal = {IEEE Transactions on Control Systems Technology},
year = {2025},
volume = {33},
number = {5},
pages = {1953--1959},
url = {https://ieeexplore.ieee.org/document/10982112},
doi = {10.1109/TCST.2025.3561059}
}
-
Multi-Scale Feature Fusion Transformer with Hybrid Attention for Insulator Defect Detection
Jiangjiao Xu, Haiyu Liao, Changjun Jiang, Dongdong Li, Ke Li
IEEE Trans. Instrumentation and Measurement (TIM), 74: 3539813, 2025
10.1109/TIM.2025.3568984
Abs | PDF | BiB | Cited by 7
In recent years, smart grid systems have rapidly advanced through the integration of sophisticated artificial intelligence (AI) technologies, leading to significant improvements in intelligence, efficiency, and sustainability. Insulators play a crucial role in the smart grid domain, yet accurately detecting small target defects remains a challenge for current detection technologies. This paper introduces a Multi-Scale Feature Fusion Transformer (MSFFT) model specifically developed to address the insufficient feature extraction capabilities in recognizing small insulator defects. This model employs a unique multi-scale feature fusion method combined with an enhanced transformer framework, enabling it to capture and process information across various scales. To further improve detection accuracy, the model incorporates a novel multi-head mixture attention mechanism. This mechanism optimizes the model’s focus on critical areas by balancing attention between fine-grained details and broader contextual features. Furthermore, this study refines the Insulator Defect Detection dataset by including annotations in both You Only Look Once (YOLO) and Common Objects in Context (COCO) formats, covering images of normal, damaged, and flashover insulators. The MSFFT model’s performance was compared with advanced insulator detection models, with experimental results demonstrating its impressive capabilities, achieving a mean average precision (mAP50:95 ) of 80.9% and an average recall (AR50:95) of 87.4%. These results underscore the model’s state-of-the-art performance and exceptional precision in detecting the smallest defects in insulators.
@article{XuLJLL25,
author = {Jiangjiao Xu and
Haiyu Liao and
Ke Li and
Changjun Jiang and
Dongdong Li},
title = {Multiscale Feature Fusion Transformer With Hybrid Attention for Insulator Defect Detection},
journal = {IEEE Transactions on Instrumentation and Measurement},
year = {2025},
volume = {74},
pages = {1--13},
url = {https://ieeexplore.ieee.org/stamp/stamp.jsp?arnumber=11000342},
doi = {10.1109/TIM.2025.3568984}
}
-
Mutual Knowledge Distillation based Personalized Federated Learning for Smart Edge Computing
Siwei Zheng, Jia Hu, Geyong Min, Ke Li
IEEE Trans. Consumer Electronics (TCE), 71(2): 6469–6476, 2025
10.1109/TCE.2024.3412817
Abs | PDF | BiB | Cited by 5
Federated Learning (FL) is a privacy-preserving machine learning paradigm that aims to train a global model using heterogeneous data across clients, which are typically consumer electronic devices such as smartphones, smart vehicles, and smart home appliances. As the global model may not be optimal for individual clients with unique behaviours, Personalized Federated Learning (PFL) was proposed to enable clients to adapt the global model to their specific needs and preferences. Nonetheless, due to the variance in data distributions across clients, the global model utilized in PFL may ‘catastrophically forget’ the knowledge gained in previous communication rounds, thereby leading to unstable performance. To address this challenge, we propose FedMKD, a novel PFL algorithm based on Mutual Knowledge Distillation (MKD) and elastic weight consolidation (EWC). FedMKD enhances the global model’s performance by addressing ‘catastrophic forgetting’ through EWC regularization, while enabling clients’ local models to effectively leverage the global model’s knowledge via MKD. Moreover, we apply uniform/exponential quantization methods to compress model parameters to decrease communication overheads. Experimental results demonstrate that FedMKD outperforms several key PFL baselines, FedMKD can also significantly reduce communication overhead while preserving its performance using suitable compression techniques, making it highly suitable for resource-constrained smart edge computing environment.
@article{ZhengHML25,
author = {Siwei Zheng and
Jia Hu and
Geyong Min and
Ke Li},
title = {Mutual Knowledge Distillation-Based Personalized Federated Learning for Smart Edge Computing},
journal = {IEEE Transactions on Consumer Electronics},
year = {2025},
volume = {71},
number = {2},
pages = {6469--6476},
url = {https://ieeexplore.ieee.org/abstract/document/10554553},
doi = {10.1109/TCE.2024.3412817}
}
-
Conversational Exploration of Literature Landscape with LitChat
Mingyu Huang, Shasha Zhou, Yuxuan Chen, Ke Li
Proc. of the 34th International Joint Conference on Artificial Intelligence (IJCAI'25)
Demo Paper Track, p. 11058–11061, August, 2025
10.24963/ijcai.2025/1262
Abs | PDF | BiB | Cited by 0
We are living in an era of "big literature", where the volume of digital scientific publications is growing exponentially. While offering new opportunities, this also poses challenges for understanding literature landscapes, as traditional manual reviewing is no longer feasible. Recent large language models (LLMs) have shown strong capabilities for literature comprehension, yet they are incapable of offering "comprehensive, objective, open and transparent" views desired by systematic reviews due to their limited context windows and trust issues like hallucinations. Here we present LitChat, an end-to-end, interactive and conversational literature agent that augments LLM agents with data-driven discovery tools to facilitate literature exploration. LitChat automatically interprets user queries, retrieves relevant sources, constructs knowledge graphs, and employs diverse data-mining techniques to generate evidence-based insights addressing user needs. We illustrate the effectiveness of LitChat via a case study on AI4Health, highlighting its capacity to quickly navigate the users through large-scale literature landscape with data-based evidence that is otherwise infeasible with traditional means.
@inproceedings{HuangZCL26,
author = {Mingyu Huang and
Shasha Zhou and
Yuxuan Chen and
Ke Li},
title = {Conversational Exploration of Literature Landscape with LitChat},
year = {2025},
pages = {11058--11061},
url = {https://dl.acm.org/doi/10.24963/ijcai.2025/1262},
doi = {10.24963/ijcai.2025/1262}
}
-
Rethinking Performance Analysis for Configurable Software Systems: A Case Study from a Fitness Landscape Perspective
Mingyu Huang, Peili Mao, Ke Li
Proc. of the ACM SIGSOFT International Symposium on Software Testing and Analysis (ISSTA'25), ISSTA077: 1748–1771, June, 2025
10.1145/3728954
Abs | PDF | Code | BiB | Cited by 1
Modern software systems are often highly configurable to tailor varied requirements from diverse stakeholders. Understanding the mapping between configurations and the desired performance attributes plays a fundamental role in advancing the controllability and tuning of the underlying system, yet has long been a dark hole of knowledge due to its black-box nature. While there have been previous efforts in performance analysis for these systems, they analyze the configurations as isolated data points without considering their inherent spatial relationships. This renders them incapable of interrogating many important aspects of the configuration space like local optima. In this work, we advocate a novel perspective to rethink performance analysis—modeling the configuration space as a structured “landscape”. To support this proposition, we utilized GraphFLA, an open-source, graph data mining empowered fitness landscape analysis (FLA) framework. By applying this framework to 86M benchmarked configurations from 32 running workloads of 3 real-world systems, we arrived at 6 main findings, which together constitute a holistic picture of the landscape topography that could have implications on both configuration tuning and performance modeling.
@inproceedings{HuangML25,
author = {Mingyu Huang and
Peili Mao and
Ke Li},
title = {Rethinking Performance Analysis for Configurable Software Systems: A Case Study from a Fitness Landscape Perspective},
booktitle = {Proceedings of the ACM on software engineering.},
year = {2025},
volume = {2},
number = {ISSTA},
pages = {1748--1771},
url = {https://dl.acm.org/doi/10.1145/3728954},
doi = {10.1145/3728954}
}
-
Faster Configuration Performance Bug Testing with Neural Dual-level Prioritization
Youpeng Ma, Tao Chen, Ke Li
Proc. of the 47th International Conference on Software Engineering (ICSE'25), p. 988–1000, May, 2025
10.1109/ICSE55347.2025.00201
Abs | PDF | Code | BiB | Cited by 3
As software systems become more complex and configurable, more performance problems tend to arise from the configuration designs. This has caused some configuration options to unexpectedly degrade performance which deviates from their original expectations designed by the developers. Such discrepancies, namely configuration performance bugs (CPBugs), are devastating and can be deeply hidden in the source code. Yet, efficiently testing CPBugs is difficult, not only due to the test oracle is hard to set, but also because the configuration measurement is expensive and there are simply too many possible configurations to test. As such, existing testing tools suffer from lengthy runtime or have been ineffective in detecting CPBugs when the budget is limited, compounded by inaccurate test oracle. In this paper, we seek to achieve significantly faster CPBug testing by neurally prioritizing the testing at both the configuration option and value range levels with automated oracle estimation. Our proposed tool, dubbed NDP, is a general framework that works with different heuristic generators. The idea is to leverage two neural language models: one to estimate the CPBug types that serve as the oracle while, more vitally, the other to infer the probabilities of an option being CPBug-related, based on which the options and the value ranges to be searched can be prioritized. Experiments on several widely-used systems of different versions reveal that NDP can, in general, better predict CPBug type in 87 % cases and find more CPBugs with up to $88.88 \times$ testing efficiency speedup over the state-of-the-art tools.
@inproceedings{MaCL25,
author = {Youpeng Ma and
Tao Chen and
Ke Li},
title = {Faster Configuration Performance Bug Testing with Neural Dual-Level Prioritization},
booktitle = {2025 IEEE/ACM 47th International Conference on Software Engineering (ICSE)},
year = {2025},
pages = {988--1000},
url = {https://dl.acm.org/doi/10.1109/ICSE55347.2025.00201},
doi = {10.1109/ICSE55347.2025.00201}
}
-
Bridging Sequence-Structure Alignment in RNA Foundation Models
Heng Yang, Renzhi Chen, Ke Li
Proc. of the 39th Annual AAAI Conference on Artificial Intelligence (AAAI'25), 39(20): 21929–21937, Philadelphia, PA, USA, February 25–March 4, 2025
10.1609/aaai.v39i20.35500
Abs | PDF | Code | BiB | ≈ 23.4% | Cited by 4
The alignment between RNA sequences and structures in foundation models (FMs) has yet to be thoroughly investigated. Existing FMs have struggled to establish sequence-structure alignment, hindering the seamless flow of genomic information between RNA sequences and structures. In this study, we introduce OmniGenome, an RNA FM trained to align RNA sequences with respect to secondary structures through structure-contextualized modelling. This alignment enables free and bidirectional mappings between sequences and structures by utilizing a flexible RNA modelling paradigm that supports versatile input and output modalities, i.e., sequence and/or structure as input/output. We implement RNA design and zero-shot secondary structure prediction as case studies to evaluate the Seq2Str and Str2Seq mapping capabilities of OmniGenome. Results on the EternaV2 benchmark show that OmniGenome solved 74% of puzzles, whereas existing FMs solved only up to 3% of the puzzles due to the lack of sequence-structure alignment. We leverage four comprehensive in-silico genome modelling benchmarks to evaluate performance across a diverse set of downstream genome tasks, where the results show that OmniGenome achieves state-of-the-art performance on RNA and DNA benchmarks, even without any training on DNA genomes.
@inproceedings{YangCL25,
author = {Heng Yang and
Renzhi Chen and
Ke Li},
title = {Bridging Sequence-Structure Alignment in RNA Foundation Models},
booktitle = {Proceedings of the AAAI Conference on Artificial Intelligence},
year = {2025},
volume = {39},
number = {20},
pages = {21929--21937},
url = {https://dl.acm.org/doi/10.1609/aaai.v39i20.35500},
doi = {10.1609/aaai.v39i20.35500}
}
-
On the Hyperparameter Loss Landscapes of Machine Learning Models: An Exploratory Study
Mingyu Huang, Ke Li
Proc. of the 31st SIGKDD Conference on Knowledge Discovery and Data Mining (KDD'25), p. 555–564, Toronto, ON, Canada, August 3–7, 2025
10.1145/3690624.3709229
Abs | PDF | Code | BiB | ≈ 19.0% | Cited by 0
Previous efforts on hyperparameter optimization (HPO) of machine learning (ML) models predominately focus on algorithmic advances, yet little is known about the topography of the underlying hyperparameter (HP) loss landscape, which plays a fundamental role in governing the search process of HPO. While several works have conducted fitness landscape analysis (FLA) on various ML systems, they are limited to properties of isolated landscape without interrogating the potential structural similarities among landscapes induced on different scenarios. The exploration of such similarities can provide a novel perspective for understanding the mechanism behind modern HPO methods, but has been missing. In this paper, we mapped 1,500 HP loss landscapes of 6 representative ML models on 63 datasets across different fidelity levels, with 11M+ configurations. By conducting exploratory analysis on these landscapes with fine-grained visualizations and dedicated FLA metrics, we observed a similar landscape topography across a wide range of models, datasets, and fidelities, and shed light on the mechanism behind the success of several popular methods in HPO. The artifacts associated with this paper is available at https://github.com/COLA-Laboratory/GraphFLA.
@inproceedings{HuangL25,
author = {Mingyu Huang and
Ke Li},
title = {On the Hyperparameter Loss Landscapes of Machine Learning Models: An Exploratory Study},
booktitle = {Proceedings of the 31st ACM SIGKDD Conference on Knowledge Discovery and Data Mining V.1},
year = {2025},
pages = {555--564},
url = {https://dl.acm.org/doi/10.1145/3690624.3709229},
doi = {10.1145/3690624.3709229}
}
-
FlowJD: Your Imagination can Help You Jailbreak in Visual Language Models
Xiaotian Zou, Yongkang Chen, Qianqian Han, Ke Li
Proc. of the 2025 IEEE International Conference on Multimedia and Expo (ICME'25)
10.1109/ICME59968.2025.11209893
Abs | PDF | Code | BiB | Cited by 0
Large Visual Language Models (VLMs), such as GPT-4V, have achieved impressive results in generating detailed and nuanced responses. Although researchers have proposed various benchmarks to evaluate VLM performance, they have often neglected the examination of inherent security capabilities, particularly by evaluating the logical comprehension of image information. To address this gap, this paper introduces a novel dataset, FlowJD, specifically designed to evaluate logical flowchart jailbreak capabilities in VLMs. We conduct a comprehensive evaluation on GPT-4o, GPT-4V, and seven other state-of-the-art VLMs, revealing jailbreak rates of up to 92.8%. Our findings reveal significant vulnerabilities in current VLMs concerning logical flowchart jailbreak, emphasizing the urgent need for robust and effective defenses in future VLM development.Warning: Some of the examples may be harmful!
@inproceedings{ZouCHL25,
author = {Xiaotian Zou and
Yongkang Chen and
Qianqian Han and
Ke Li},
title = {FlowJD: Your Imagination Can Help You Jailbreak in Visual Language Models},
booktitle = {2025 IEEE International Conference on Multimedia and Expo (ICME)},
year = {2025},
pages = {1--6},
url = {https://ieeexplore.ieee.org/document/11209893},
doi = {10.1109/ICME59968.2025.11209893}
}
-
An Interpretable RNA Foundation Model for Exploration Functional RNA Motifs in Plants
Haopeng Yu, Heng Yang, Wenqing Sun, Zongyun Yan, Xiaofei Yang, Huakun Zhang, Yiliang Ding, Ke Li
Nature Machine Intelligence (NMI), 6: 1616–1625, 2024
10.1038/s42256-024-00946-z
Abs | PDF | Supp | Code | Project | BiB | Cited by 30
The complex 'language' of plant RNA encodes a vast array of biological regulatory elements that orchestrate crucial aspects of plant growth, development and adaptation to environmental stresses. Recent advancements in foundation models (FMs) have demonstrated their unprecedented potential to decipher complex 'language' in biology. In this study, we introduced PlantRNA-FM, a high-performance and interpretable RNA FM specifically designed for plants. PlantRNA-FM was pretrained on an extensive dataset, integrating RNA sequences and RNA structure information from 1,124 distinct plant species. PlantRNA-FM exhibits superior performance in plant-specific downstream tasks. PlantRNA-FM achieves an F1 score of 0.974 for genic region annotation, whereas the current best-performing model achieves 0.639. Our PlantRNA-FM is empowered by our interpretable framework that facilitates the identification of biologically functional RNA sequence and structure motifs, including both RNA secondary and tertiary structure motifs across transcriptomes. Through experimental validations, we revealed translation-associated RNA motifs in plants. Our PlantRNA-FM also highlighted the importance of the position information of these functional RNA motifs in genic regions. Taken together, our PlantRNA-FM facilitates the exploration of functional RNA motifs across the complexity of transcriptomes, empowering plant scientists with capabilities for programming RNA codes in plants.
@article{YuYSYYZDL24,
author = {Haopeng Yu and
Heng Yang and
Wenqing Sun and
Zongyun Yan and
Xiaofei Yang and
Huakun Zhang and
Yiliang Ding and
Ke Li},
title = {An interpretable RNA foundation model for exploring functional RNA motifs in plants},
journal = {Nature Machine Intelligence},
year = {2024},
volume = {6},
number = {12},
pages = {1616--1625},
url = {https://doi.org/10.1038/s42256-024-00946-z},
doi = {10.1038/s42256-024-00946-z}
}
-
Quality Indicators for Preference-based Evolutionary Multi-objective Optimization Using a Reference Point: A Review and Analysis
Ryoji Tanabe, Ke Li
IEEE Trans. Evolutionary Computation (TEVC), 28(6): 1575–1589, 2024
10.1109/TEVC.2023.3319009
Abs | PDF | Supp | BiB | Cited by 4
Some quality indicators have been proposed for benchmarking preference-based evolutionary multi-objective optimization algorithms using a reference point. Although a systematic review and analysis of the quality indicators are helpful for both benchmarking and practical decision-making, neither has been conducted. In this context, first, this paper reviews existing regions of interest and quality indicators for preference-based evolutionary multi-objective optimization using the reference point. We point out that each quality indicator was designed for a different region of interest. Then, this paper investigates the properties of the quality indicators. We demonstrate that an achievement scalarizing function value is not always consistent with the distance from a solution to the reference point in the objective space. We observe that the regions of interest can be significantly different depending on the position of the reference point and the shape of the Pareto front. We identify undesirable properties of some quality indicators. We also show that the ranking of preference-based evolutionary multi-objective optimization algorithms depends on the choice of quality indicators.
@article{TanabeL23,
author = {Ryoji Tanabe and
Ke Li},
title = {Quality Indicators for Preference-Based Evolutionary Multiobjective Optimization Using a Reference Point: A Review and Analysis},
journal = {IEEE Transactions on Evolutionary Computation},
year = {2024},
volume = {28},
number = {6},
pages = {1575--1589},
url = {https://ieeexplore.ieee.org/document/10268021},
doi = {10.1109/TEVC.2023.3319009}
}
-
Solving Expensive Optimization Problems in Dynamic Environments with Meta-Learning
Huan Zhang, Jinliang Ding, Liang Feng, Kay Chen Tan, Ke Li
IEEE Trans. Cybernetics (TCYB), 52(12): 7430–7442, 2024
10.1109/TCYB.2024.3443396
Abs | PDF | BiB | Cited by 13
Dynamic environments pose great challenges for expensive optimization problems, as the objective functions of these problems change over time and thus require remarkable computational resources to track the optimal solutions. Although data-driven evolutionary optimization and Bayesian optimization (BO) approaches have shown promise in solving expensive optimization problems in static environments, the attempts to develop such approaches in dynamic environments remain rarely explored. In this article, we propose a simple yet effective meta-learning-based optimization framework for solving the expensive dynamic optimization problems. This framework is flexible, allowing any off-the-shelf continuously differentiable surrogate model to be used in a plug-in manner, either in data-driven evolutionary optimization or BO approaches. In particular, the framework consists of two unique components: 1) the meta-learning component, in which a gradient-based meta-learning approach is adopted to learn experience (effective model parameters) across different dynamics along the optimization process and 2) the adaptation component, where the learned experience (model parameters) is used as the initial parameters for fast adaptation in the dynamic environment based on few shot samples. By doing so, the optimization process is able to quickly initiate the search in a new environment within a strictly restricted computational budget. Experiments demonstrate the effectiveness of the proposed algorithm framework compared to several state-of-the-art algorithms on common benchmark test problems under different dynamic characteristics.
@article{ZhangDFTL24,
author = {Huan Zhang and
Jinliang Ding and
Liang Feng and
Kay Chen Tan and
Ke Li},
title = {Solving Expensive Optimization Problems in Dynamic Environments With Meta-Learning},
journal = {IEEE Transactions on Cybernetics},
year = {2024},
volume = {54},
number = {12},
pages = {7430--7442},
url = {https://ieeexplore.ieee.org/abstract/document/10644136},
doi = {10.1109/TCYB.2024.3443396}
}
-
A Data-Driven Evolutionary Transfer Optimization for Expensive Problems in Dynamic Environments
Ke Li*, Renzhi Chen*, Xin Yao
IEEE Trans. Evolutionary Computation (TEVC), 28(5): 1396–1411, 2024
10.1109/TEVC.2023.3307244
Abs | PDF | Supp | Code | BiB | Cited by 23
Many real-world problems are computationally costly and the objective functions evolve over time. Data-driven, a.k.a. surrogate-assisted, evolutionary optimization has been recognized as an effective approach to tackle expensive black-box optimization problems in a static environment whereas it has rarely been studied under dynamic environments. This paper proposes a simple yet effective transfer learning framework to empower data-driven evolutionary optimization to solve expensive dynamic optimization problems. Specifically, a hierarchical multi-output Gaussian process is proposed to capture the correlation among data collected from different time steps with a linearly increased number of hyperparameters. Furthermore, an adaptive source task selection along with a bespoke warm staring initialization mechanisms are proposed to better leverage the knowledge extracted from previous optimization processes. By doing so, the data-driven evolutionary optimization can jump start the optimization in the new environment with a very limited computational budget. Experiments on synthetic benchmark test problems and a real-world case study demonstrate the effectiveness of our proposed algorithm in comparison with nine state-of-the-art peer algorithms.
@article{LiCY23,
author = {Ke Li and
Renzhi Chen and
Xin Yao},
title = {A Data-Driven Evolutionary Transfer Optimization for Expensive Problems in Dynamic Environments},
journal = {IEEE Transactions on Evolutionary Computation},
year = {2024},
volume = {28},
number = {5},
pages = {1396--1411},
url = {https://ieeexplore.ieee.org/document/10225543},
doi = {10.1109/TEVC.2023.3307244}
}
-
A Many-Objective Evolutionary Algorithm Based on Interaction Force and Hybrid Optimization Mechanism
Lei Yang, Jiale Cao, Kangshun Li, Yuanye Zhang, Rui Xu, Ke Li
Swarm and Evolutionary Computation (SWEVO), 90: 101667
10.1016/j.swevo.2024.101667
Abs | PDF | BiB | Cited by 10
In many-objective optimization, both convergence and diversity are equally important. However, in high-dimensional spaces, traditional decomposition-based many-objective evolutionary algorithms struggle to ensure population diversity. Conversely, traditional Pareto dominance-based many-objective evolutionary algorithms face challenges in ensuring population convergence. In this paper, we propose a novel many-objective evolutionary algorithm based on interaction force and hybrid optimization mechanism (MaOEAIH) for effectively addressing the difficulty in balancing convergence and diversity. First, we use the concept of interaction force to simulate the convergence (akin to gravity) and diversity (repulsion) of the population. Subsequently, we design an optimization mechanism that combines decomposition and Pareto dominance to enhance the convergence and diversity of the population separately. Simultaneously, to eliminate dominance resistance solutions, we propose a quartile method based on boundary solutions. Additionally, Random perturbations are also introduced to certain individuals within the population to facilitate their escape from local optima. MaOEAIH is compared with some state-of-the-art algorithms on 31 well-known test problems with 3-15 objectives. The experimental results show that, compared to other algorithms, MaOEAIH not only obtains solution sets of higher quality when dealing with different types of many-objective optimization problems, but also effectively addresses key challenges including insufficient selection pressure, difficulty balancing convergence and diversity, and susceptibility to population entrapment in local optima within many-objective optimization scenarios.
@article{YangCLZXL24,
author = {Lei Yang and
Jiale Cao and
Kangshun Li and
Yuanye Zhang and
Rui Xu and
Ke Li},
title = {A many-objective evolutionary algorithm based on interaction force and hybrid optimization mechanism},
journal = {Swarm and Evolutionary Computation},
year = {2024},
volume = {90},
pages = {101667},
url = {https://www.sciencedirect.com/science/article/abs/pii/S2210650224002050},
doi = {10.1016/j.swevo.2024.101667}
}
-
iM-Seeker: A Webserver for DNA I-motifs Prediction and Scoring via Automated Machine Learning
Haopeng Yu, Fan Li, Bibo Yang, Yiman Qi, Dilek Guneri, Wenqian Chen, Zoe Waller, Ke Li, Yiliang Ding
Nucleic Acids Research (NAR), 52(W1): W19-W28, 2024
10.1093/nar/gkae315
Abs | PDF | Code | BiB | Cited by 18
DNA, beyond its canonical B-form double helix, adopts various alternative conformations, among which the i-motif, emerging in cytosine-rich sequences under acidic conditions, holds significant biological implications in transcription modulation and telomere biology. Despite recognizing the crucial role of i-motifs, predictive software for i-motif forming sequences has been limited. Addressing this gap, we introduce 'iM-Seeker', an innovative computational platform designed for the prediction and evaluation of i-motifs. iM-Seeker exhibits the capability to identify potential i-motifs within DNA segments or entire genomes, calculating stability scores for each predicted i-motif based on parameters such as the cytosine tracts number, loop lengths, and sequence composition. Furthermore, the webserver leverages automated machine learning (AutoML) to effortlessly fine-tune the optimal i-motif scoring model, incorporating user-supplied experimental data and customised features. As an advanced, versatile approach, 'iM-Seeker' promises to advance genomic research, highlighting the potential of i-motifs in cell biology and therapeutic applications. The webserver is freely available at https://im-seeker.org.
@article{YuLYQGCWLD24,
author = {Haopeng Yu and
Fan Li and
Bibo Yang and
Yiman Qi and
Dilek Guneri and
Wenqian Chen and
Zoë A E Waller and
Ke Li and
Yiliang Ding},
title = {iM-Seeker: a webserver for DNA i-motifs prediction and scoring via automated machine learning},
journal = {Nucleic Acids Research},
year = {2024},
volume = {52},
number = {W1},
pages = {W19--W28},
url = {https://academic.oup.com/nar/article/52/W1/W19/7659304},
doi = {10.1093/nar/gkae315}
}
-
Multi-Output Framework for Time-Series Forecasting in Smart Grid Meets Data Scarcity
Jiangjiao Xu, Ke Li, Dongdong Li
IEEE Trans. Industrial Informatics (TII), 20(9): 11202–11212, 2024
10.1109/TII.2024.3396347
Abs | PDF | BiB | Cited by 11
Sensor technology has become increasingly prevalent in various domains of human life. However, the collected data often contains missing values to varying degrees. Moreover, obtaining sufficient historical data, particularly for smart grid data forecasting in isolated networks, is often challenging. These data deficiencies can negatively impact the forecasting accuracy of deep-learning models, consequently affecting the operational performance of microgrids. To address these challenges, this article introduces a multioutput learning framework based on the multioutput Gaussian process (MOGP) model. This framework aims to achieve data imputation and prediction by leveraging the correlation between tasks simultaneously, even with limited data availability. To assess the effectiveness of the proposed method, experiments are conducted on three types of data. The empirical results demonstrate that the MOGP model outperforms two alternative techniques in terms of imputation and forecasting performance across all cases. Furthermore, to mitigate computational complexity, a novel kernel approximation method based on random Fourier features is proposed. The experimental results validate the effectiveness of this approach, as it significantly reduces computational complexity while maintaining satisfactory performance levels.
@article{XuLL24a,
author = {Jiangjiao Xu and
Ke Li and
Dongdong Li},
title = {Multioutput Framework for Time-Series Forecasting in Smart Grid Meets Data Scarcity},
journal = {IEEE Transactions on Industrial Informatics},
year = {2024},
volume = {20},
number = {9},
pages = {11202--11212},
url = {https://ieeexplore.ieee.org/document/10539288},
doi = {10.1109/TII.2024.3396347}
}
-
A Knee Point Driven Evolutionary Algorithm for Multi-Objective Bilevel Optimization
Jiaxin Chen, Jinliang Ding, Ke Li, Kay Chen Tan, Tianyou Chai
IEEE Trans. Cybernetics (TCYB), 54(7): 4177–4189, 2024
10.1109/TCYB.2024.3377272
Abs | PDF | Code | BiB | Cited by 19
Bilevel optimization is a special type of optimization in which one problem is embedded within another. The bilevel optimization problem (BLOP) of which both levels are multiobjective functions is usually called the multiobjective BLOP (MBLOP). The expensive computation and nested features make it challenging to solve. Most existing studies look for complete lower-level solutions for every upper-level variable. However, not every lower-level solution will participate in the bilevel Pareto-optimal front. Under a limited computational budget, instead of wasting resources to find complete lower-level solutions that may not be in the feasible region or inducible region of the MBLOP, it is better to concentrate on finding the solutions with better performance. Bearing these considerations in mind, we propose a multiobjective bilevel optimization solving routine combined with a knee point driven algorithm. Specifically, the proposed algorithm aims to quickly find feasible solutions considering the lower-level constraints in the first stage and then concentrates the computational resources on finding solutions with better performance. Besides, we develop several multiobjective bilevel test problems with different properties, such as scalable, deceptive, convexity, and (dis)continuous. Finally, the performance of the algorithm is validated on a practical petroleum refining bilevel problem, which involves a multiobjective environmental regulation problem and a petroleum refining operational problem. Comprehensive experiments fully demonstrate the effectiveness of our presented algorithm in solving MBLOPs.
@article{ChenDLTC24,
author = {Jiaxin Chen and
Jinliang Ding and
Ke Li and
Kay Chen Tan and
Tianyou Chai},
title = {A Knee Point Driven Evolutionary Algorithm for Multiobjective Bilevel Optimization},
journal = {IEEE Transactions on Cybernetics},
year = {2024},
volume = {54},
number = {7},
pages = {4177--4189},
url = {https://ieeexplore.ieee.org/document/10496817},
doi = {10.1109/TCYB.2024.3377272}
}
-
An Automated Few-Shot Learning for Time Series Forecasting in Smart Grid Under Data Scarcity
Jiangjiao Xu, Ke Li, D. Li
IEEE Trans. Artificial Intelligence (TAI), 5(6): 2482–2492, 2024
10.1109/TAI.2024.3358795
Abs | PDF | Code | BiB | Cited by 16
Micro-grid can improve greenhouse gas emissions and reduce operational costs. To forecast both energy generation and load demand, time series prediction has been a key tool in real-time control and optimization. Developing an adequate predictive model is difficult when there is a lack of historical data. Moreover, hyperparameters have a tangible impact on the performance of machine learning models. Bearing these considerations in mind, this paper develops a BiLO-Auto-TSF/ML framework that automates the optimal design of a few-shot learning pipeline from a bi-level programming perspective. Specifically, a lower-level meta-learner helps mitigate the small data challenge, whereas an upper-level optimization optimizes both hyperparameters for lower- and upper-level learners. Note that the proposed framework is designed to accommodate a wide range of machine learning methods, allowing for easy integration through a plug-in mechanism. Comprehensive experiments demonstrate the effectiveness of our proposed BiLO-Auto- TSF/ML to search for a high-performance few-shot learning pipeline for various energy sources.
@article{XuLL24b,
author = {Jiangjiao Xu and
Ke Li and
Dongdong Li},
title = {An Automated Few-Shot Learning for Time-Series Forecasting in Smart Grid Under Data Scarcity},
journal = {IEEE Transactions on Artificial Intelligence},
year = {2024},
volume = {5},
number = {6},
pages = {2482--2492},
url = {https://ieeexplore.ieee.org/abstract/document/10414362},
doi = {10.1109/TAI.2024.3358795}
}
-
Evolutionary Bi-level Optimization via Multi-objective Transformation Based Lower Level Search
Lei Chen, Hai-Lin Liu, Ke Li, Kay Chen Tan
IEEE Trans. Evolutionary Computation (TEVC), 28(3): 733–747, 2024
10.1109/TEVC.2023.3236455
Abs | PDF | BiB | Cited by 15
Nested evolutionary algorithms (EAs) have been regarded as very promising tools for bi-level optimization. Due to the nested structure, the upper level population evaluation requires a set of complete lower level optimizations, thereby reducing the efficiency and practicability of EA methods. In this paper, a multi-objective transformation-based evolutionary algorithm (MOTEA) is proposed to perform multiple lower level optimizations in a parallel and collaborative manner. Specifically, the corresponding multiple lower level optimizations for each generation of the upper level population evaluation are transformed into locating a set of Pareto optimal solutions of a constructed multi-objective optimization problem. By utilizing the built-in implicit parallelism of evolutionary multi-objective optimization, multiple lower level problems can thus be optimized in parallel. Within one multi-objective search population, the collaboration among the parallel lower level optimization can be realized by exploiting and utilizing the implicit similarities among them for better efficiency. The effectiveness and efficiency of the proposed MOTEA are verified by comparing it with four state-of-the-art evolutionary bi-level optimization algorithms on two sets of popular bi-level optimization benchmark test problems and three application problems.
@article{ChenLLT24,
author = {Lei Chen and
Hai-Lin Liu and
Ke Li and
Kay Chen Tan},
title = {Evolutionary Bilevel Optimization via Multiobjective Transformation-Based Lower-Level Search},
journal = {IEEE Transactions on Evolutionary Computation},
year = {2024},
volume = {28},
number = {3},
pages = {733--747},
url = {https://ieeexplore.ieee.org/document/10016238},
doi = {10.1109/TEVC.2023.3236455}
}
-
Direct Preference-Based Evolutionary Multi-Objective Optimization with Dueling Bandits
Tian Huang, Shengbo Wang, Ke Li
Proc. of the 38th Annual Conference on Neural Information Processing Systems (NeurIPS'24)
PDF | Code | BiB | ≈ 25.8%
-
MP-RNA: Unleashing Multi-species RNA Foundation Model via Calibrated Secondary Structure Prediction
Heng Yang, Ke Li
Findings of the 2024 Conference on Empirical Methods in Natural Language Processing (EMNLP'24)
10.18653/v1/2024.findings-emnlp.304
Abs | PDF | Code | BiB | ≈ 16.9% | Cited by 6
RNA foundation models (FMs) have been extensively used to interpret genomic sequences and address a wide range of in-silico genomic tasks.However, current RNA FMs often overlook the incorporation of secondary structures in the pretraining of FMs, which impedes the effectiveness in various genomic tasks.To address this problem, we leverage filtered highfidelity structure annotations for structure pretraining to enhance the modeling ability of FMs in single nucleotide resolution tasks.Experimental evaluations across four comprehensive genomic benchmarks demonstrate that our FM (MP-RNA) consistently outperforms existing RNA FMs, achieving a 40% improvement in RNA secondary structure prediction and obtaining top-tier results on DNA genomic benchmarks even though it has not been pretrained on any DNA genome.We release the code and tutorials 1 and models to encourage further research to bridge the gap between in-silico predictions and biological reality.
@article{Yang2024,
author = {Heng Yang and
Ke Li},
title = {MP-RNA: Unleashing Multi-species RNA Foundation Model via Calibrated Secondary Structure Prediction},
journal = {Findings of the Association for Computational Linguistics: EMNLP 2024},
year = {2024},
pages = {5278--5296},
url = {https://aclanthology.org/2024.findings-emnlp.304/},
doi = {10.18653/v1/2024.findings-emnlp.304}
}
-
The Best Defense is Attack: Repairing Semantics in Textual Adversarial Examples
Heng Yang, Ke Li
Proc. of the 2024 Conference on Empirical Methods in Natural Language Processing (EMNLP'24)
10.18653/v1/2024.emnlp-main.481
Abs | PDF | BiB | ≈ 20.8% | Cited by 3
Recent studies have revealed the vulnerability of pre-trained language models to adversarial attacks.Adversarial defense techniques have been proposed to reconstruct adversarial examples within feature or text spaces.However, these methods struggle to effectively repair the semantics in adversarial examples, resulting in unsatisfactory defense performance.To repair the semantics in adversarial examples, we introduce a novel approach named Reactive Perturbation Defocusing (RAPID), which employs an adversarial detector to identify the fake labels of adversarial examples and leverages adversarial attackers to repair the semantics in adversarial examples.Our extensive experimental results, conducted on four public datasets, demonstrate the consistent effectiveness of RAPID in various adversarial attack scenarios.For easy evaluation, we provide a click-to-run demo of RAPID at https://tinyurl.com/22ercuf8.
@article{Yang2024,
author = {Heng Yang and
Ke Li},
title = {The Best Defense is Attack: Repairing Semantics in Textual Adversarial Examples},
journal = {Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing},
year = {2024},
pages = {8439--8457},
url = {https://aclanthology.org/2024.emnlp-main.481/},
doi = {10.18653/v1/2024.emnlp-main.481}
}
-
OpenTOS: Open-source System for Transfer Learning Bayesian Optimization
Peili Mao, Ke Li
Proc. of the 33rd ACM International Conference on Information and Knowledge Management (CIKM'24) Demo Paper track
10.1145/3627673.3679225
Abs | PDF | BiB | ≈ 43.2% | Cited by 0
In recent years, many studies successfully integrated transfer learning techniques to improve the performance of Bayesian optimization. However, these advanced methods have not been widely adopted in real-world applications due to their inherent complexity and challenges in re-implementation and reproducibility. In this work, we introduce OpenTOS, an open-source system designed for transfer learning in Bayesian optimization. OpenTOS introduces a new implementation paradigm for these methods, allowing users to build different algorithms by choosing algorithmic components, similar to assembling LEGO blocks. Additionally, OpenTOS provides robust data management for supporting transfer learning with data from various sources. We also developed a web interface that allows for interactive building, analysis, and visualization of the optimization process. Powered by LLM, this interface offers a conversational experience, allowing users to interact with the system through natural language dialogue. OpenTOS is available as open-source on https://github.com/COLA-Laboratory/TransOPTGitHub .
@article{Mao2024,
author = {Peili Mao and
Ke Li},
title = {OpenTOS: Open-source System for Transfer Learning Bayesian Optimization},
journal = {Proceedings of the 33rd ACM International Conference on Information and Knowledge Management},
year = {2024},
pages = {5254--5259},
url = {https://dl.acm.org/doi/10.1145/3627673.3679225},
doi = {10.1145/3627673.3679225}
}
-
RNAInvBench: Benchmark for the RNA Inverse Design Problem
Jack Cole, Fan Li, Liwen Wu, Krasmira Tsaneva-Atanasova, Ke Li
Proc. of the 2024 ICML AI for Science Workshop
PDF | Code | BiB
-
Evolutionary Multi-Objective Optimization for Contextual Adversarial Example Generation
Shasha Zhou, Mingyu Huang, Yanan Sun, Ke Li
Proc. of the 2024 ACM International Conference on the Foundations of Software Engineering (FSE'24), 101: 2285–2308, July, 2024
10.1145/3660808
Abs | PDF | Code | BiB | ≈ 21.0% | Cited by 12
The emergence of the ‘code naturalness’ concept, which suggests that software code shares statistical properties with natural language, paves the way for deep neural networks (DNNs) in software engineering (SE). However, DNNs can be vulnerable to certain human imperceptible variations in the input, known as adversarial examples (AEs), which could lead to adverse model performance. Numerous attack strategies have been proposed to generate AEs in the context of computer vision and natural language processing, but the same is less true for source code of programming languages in SE. One of the challenges is derived from various constraints including syntactic, semantics and minimal modification ratio. These constraints, however, are subjective and can be conflicting with the purpose of fooling DNNs. This paper develops a multi-objective adversarial attack method (dubbed MOAA ), a tailored NSGA-II, a powerful evolutionary multi-objective (EMO) algorithm, integrated with CodeT5 to generate high-quality AEs based on contextual information of the original code snippet. Experiments on 5 source code tasks with 10 datasets of 6 different programming languages show that our approach can generate a diverse set of high-quality AEs with promising transferability. In addition, using our AEs, for the first time, we provide insights into the internal behavior of pre-trained models.
@article{Zhou2024,
author = {Shasha Zhou and
Mingyu Huang and
Yanan Sun and
Ke Li},
title = {Evolutionary Multi-objective Optimization for Contextual Adversarial Example Generation},
journal = {Proceedings of the ACM on software engineering.},
year = {2024},
volume = {1},
number = {FSE},
pages = {2285--2308},
url = {https://dl.acm.org/doi/10.1145/3660808},
doi = {10.1145/3660808}
}
-
Constrained Bayesian Optimization Under Partial Observations: Balanced Improvements and Provable Convergence
Shengbo Wang, Ke Li
Proc. of the 38th Annual AAAI Conference on Artificial Intelligence (AAAI'24), 38(14): 15607-15615, February, 2024
10.1609/aaai.v38i14.29488
Abs | PDF | Code | BiB | ≈ 23.5% | Cited by 12
The partially observable constrained optimization problems (POCOPs) impede data-driven optimization techniques since an infeasible solution of POCOPs can provide little information about the objective as well as the constraints. We endeavor to design an efficient and provable method for expensive POCOPs under the framework of constrained Bayesian optimization. Our method consists of two key components. Firstly, we present an improved design of the acquisition functions that introduce balanced exploration during optimization. We rigorously study the convergence properties of this design to demonstrate its effectiveness. Secondly, we propose Gaussian processes embedding different likelihoods as the surrogate model for partially observable constraints. This model leads to a more accurate representation of the feasible regions compared to traditional classification-based models. Our proposed method is empirically studied on both synthetic and real-world problems. The results demonstrate the competitiveness of our method for solving POCOPs.
@inproceedings{WangL23,
author = {Shengbo Wang and
Ke Li},
title = {Constrained Bayesian Optimization under Partial Observations: Balanced Improvements and Provable Convergence},
booktitle = {Proceedings of the AAAI Conference on Artificial Intelligence},
year = {2024},
volume = {38},
number = {14},
pages = {15607--15615},
url = {https://ojs.aaai.org/index.php/AAAI/article/view/29488},
doi = {10.1609/aaai.v38i14.29488}
}
-
Modeling Aspect Sentiment Coherency via Local Sentiment Aggregation
Heng Yang, Ke Li
Findings of the 18th Conference of the European Chapter of the Association for Computational Linguistics (EACL'24), p. 182-195, March, 2024.
10.18653/v1/2024.findings-eacl.13
Abs | PDF | BiB | Cited by 0
Aspect sentiment coherency is an intriguing yet underexplored topic in the field of aspectbased sentiment classification.This concept reflects the common pattern where adjacent aspects often share similar sentiments.Despite its prevalence, current studies have not fully recognized the potential of modeling aspect sentiment coherency, including its implications in adversarial defense.To model aspect sentiment coherency, we propose a novel local sentiment aggregation (LSA) paradigm based on constructing a differential-weighted sentiment aggregation window.We have rigorously evaluated our model through experiments, and the results affirm the proficiency of LSA in terms of aspect coherency prediction and aspect sentiment classification.For instance, it outperforms existing models and achieves stateof-the-art sentiment classification performance across five public datasets.Furthermore, we demonstrate the promising ability of LSA in ABSC adversarial defense, thanks to its sentiment coherency modeling.To encourage further exploration and application of this concept, we have made our code publicly accessible.This will provide researchers with a valuable tool to delve into sentiment coherency modeling in future research.
@article{Yang2024,
author = {Heng Yang and
Ke Li},
title = {Modeling Aspect Sentiment Coherency via Local Sentiment Aggregation},
journal = {Findings of the Association for Computational Linguistics: EACL 2024},
year = {2024},
pages = {182--195},
url = {https://aclanthology.org/2024.findings-eacl.13},
doi = {10.18653/v1/2024.findings-eacl.13}
}
-
Multidimensional Resource Fragmentation-Aware Virtual Network Embedding for IoT Applications in MEC Networks
Yingying Guan, Qingyang Song, Weijing Qi, Lei Guo, Ke Li, Abbas Jamalipour
IEEE Internet of Things Journal (IoTJ), 10(24): 22223–22232, 2023
10.1109/JIOT.2023.3304976
Abs | PDF | BiB | Cited by 10
The proliferation of Internet of Things (IoT) applications has led to the interconnection of multiaccess edge computing (MEC) systems through metro optical networks. To cater to these diverse applications, network slicing has become a popular tool for creating specialized virtual networks. However, the uneven utilization of multidimensional resources can result in resource fragmentation, thereby reducing the utilization of limited edge resources. This article focuses on mitigating multidimensional resource fragmentation in virtual network embedding (VNE) to maximize the profit of the infrastructure provider (InP). The problem is converted into a bilevel optimization problem, taking into account the interdependence between virtual node embedding and virtual link embedding. To solve this problem, we propose a nested bilevel VNE approach named BiVNE. BiVNE leverages an ant colony system (ACS) algorithm for the upper layer problem and utilizes the Dijkstra algorithm and an exact-fit spectrum slot assignment method for the lower layer problem. Evaluation results demonstrate that BiVNE can greatly improve the profit of the InP by increasing the acceptance ratio and avoiding resource fragmentation simultaneously.
@article{GuanSQLGJ23,
author = {Yingying Guan and
Qingyang Song and
Weijing Qi and
Lei Guo and
Ke Li and
Abbas Jamalipour},
title = {Multidimensional Resource Fragmentation-Aware Virtual Network Embedding for IoT Applications in MEC Networks},
journal = {IEEE Internet of Things Journal},
year = {2023},
volume = {10},
number = {24},
pages = {22223--22232},
url = {https://ieeexplore.ieee.org/document/10217060},
doi = {10.1109/JIOT.2023.3304976}
}
-
Interactive Evolutionary Multi-Objective Optimization via Learning-to-Rank
Ke Li, Guiyu Lai, Xin Yao
IEEE Trans. Evolutionary Computation (TEVC), 27(4): 749–763, 2023
10.1109/TEVC.2023.3234269
Abs | PDF | Supp | BiB | Cited by 25
In practical multicriterion decision making, it is cumbersome if a decision maker (DM) is asked to choose among a set of tradeoff alternatives covering the whole Pareto-optimal front. This is a paradox in conventional evolutionary multiobjective optimization (EMO) that always aim to achieve a well balance between convergence and diversity. In essence, the ultimate goal of multiobjective optimization is to help a DM identify solution(s) of interest (SOI) achieving satisfactory tradeoffs among multiple conflicting criteria. Bearing this in mind, this article develops a framework for designing preference-based EMO algorithms to find SOI in an interactive manner. Its core idea is to involve human in the loop of EMO. After every several iterations, the DM is invited to elicit her feedback with regard to a couple of incumbent candidates. By collecting such information, her preference is progressively learned by a learning-to-rank neural network and then applied to guide the baseline EMO algorithm. Note that this framework is so general that any existing EMO algorithm can be applied in a plug-in manner. Experiments on 48 benchmark test problems with up to ten objectives and a real-world multiobjective robot control problem fully demonstrate the effectiveness of our proposed algorithms for finding SOI.
@article{LiLY23,
author = {Ke Li and
Guiyu Lai and
Xin Yao},
title = {Interactive Evolutionary Multiobjective Optimization via Learning to Rank},
journal = {IEEE Transactions on Evolutionary Computation},
year = {2023},
volume = {27},
number = {4},
pages = {749--763},
url = {https://ieeexplore.ieee.org/document/10015671},
doi = {10.1109/TEVC.2023.3234269}
}
-
Efficient Spectral Graph Convolutional Network Deployment on Memristive Crossbars
Bo Lyu, Maher Hamdi, Yin Yang, Yuting Cao, Zheng Yan, Ke Li, Shiping Wen, Tingwen Huang
IEEE Trans. Emerging Topics in Computational Intelligence (TETCI), 7(2): 415–425, 2023
10.1109/TETCI.2022.3210998
Abs | PDF | BiB | Cited by 28
Graph Neural Networks (GNNs) have attracted increasing research interest for their remarkable capability to model graph-structured knowledge. However, GNNs suffer from intensive data exchange and poor data locality, which will cause critical performance and energy bottlenecks under conventional complementary metal oxide semiconductor (CMOS)-based von-Neumann computing architectures (graphics processing unit (GPU), central processing unit (CPU)) for the “Memory Wall” issue. Fortunately, memristive crossbar-based computation has emerged as one of the most promising neuromorphic computing architectures, which has been widely studied as the computing platform for convolutional neural network (CNNs), recurrent neural network (RNNs), spiking neural network (SNNs), etc. This paper proposes the deployment of spectral graph convolutional networks (GCNs) on memristive crossbars. Further, based on the structure of GCNs (extremely high sparsity and unbalanced non-zero data distribution) and the neuromorphic characteristics of memristive crossbar circuit, we propose the acceleration method that consists of Sparse Laplace Matrix Reordering and Diagonal Block Matrix Multiplication. The simulated experiment on memristor crossbars achieves 90.3% overall accuracy on the supervised learning graph dataset (QM7), and compared with the original computation, the proposed acceleration computing framework (with half-size diagonal blocks) achieves a 27.3% reduction of memristor number. Additionally, on the unsupervised learning dataset (Karate club), our method shows no loss of accuracy with half-size diagonal block mapping and reaches a 32.2% reduction of memristor number.
@article{LyuHYCYLSH22,
author = {Bo Lyu and
Maher Hamdi and
Yin Yang and
Yuting Cao and
Zheng Yan and
Ke Li and
Shiping Wen and
Tingwen Huang},
title = {Efficient Spectral Graph Convolutional Network Deployment on Memristive Crossbars},
journal = {IEEE Transactions on Emerging Topics in Computational Intelligence},
year = {2023},
volume = {7},
number = {2},
pages = {415--425},
url = {https://ieeexplore.ieee.org/document/9918532/},
doi = {10.1109/TETCI.2022.3210998}
}
-
Neural Architecture Search for Portrait Parsing
Bo Lyu, Yin Yang, Shiping Wen, Tingwen Huang, Ke Li
IEEE Trans. Neural Networks and Learning Systems (TNNLS), 34(3): 1112–1121, 2023
10.1109/TNNLS.2021.3104872
Abs | PDF | BiB | Cited by 21
This work proposes a neural architecture search (NAS) method for portrait parsing, which is a novel up-level task based on portrait segmentation and face labeling. Recently, NAS has become an effective method in terms of automatic machine learning. However, remarkable achievements have been made only in image classification and natural language processing (NLP) areas. Meanwhile, state-of-the-art portrait segmentation and face labeling approaches are all manually designed, but few models reach a tradeoff between efficiency and performance. Thus, we are extremely interested in improving existing NAS methods for dense-per-pixel prediction tasks on portrait datasets. To achieve that, we resort to a cell-based encoder-decoder architecture with an elaborate design of connectivity structure and searching space. As a result, we achieve state-of-the-art performance on three portrait tasks, including 96.8% MIOU on EG1800 (portrait segmentation), 91.2% overall F1 -score on HELEN (face labeling), and 95.1% overall F1 -score on CelebAMask-HQ (portrait parsing) with only 2.29M model parameters. That is, our approach compares favorably with all previous works on portrait datasets. More crucially, we empirically prove that even a fundamental encoder-decoder architecture may reach an outstanding result on the aforementioned tasks with the help of the innovative approach of NAS. To the best of our knowledge, our work is also the first to report the success of applying NAS on these portrait tasks.
@article{LyuYWHL23,
author = {Bo Lyu and
Yin Yang and
Shiping Wen and
Tingwen Huang and
Ke Li},
title = {Neural Architecture Search for Portrait Parsing},
journal = {IEEE Transactions on Neural Networks and Learning Systems},
year = {2023},
volume = {34},
number = {3},
pages = {1112--1121},
url = {https://ieeexplore.ieee.org/document/9518382},
doi = {10.1109/TNNLS.2021.3104872}
}
-
Batched Data-Driven Evolutionary Multi-Objective Optimization Based on Manifold Interpolation
Ke Li*, Renzhi Chen*
IEEE Trans. Evolutionary Computation (TEVC), 27(1): 126–140, 2023
10.1109/TEVC.2022.3162993
Abs | PDF | Supp | BiB | Cited by 33
Multiobjective optimization problems are ubiquitous in real-world science, engineering, and design optimization problems. It is not uncommon that the objective functions are as a black box, the evaluation of which usually involve time-consuming and/or costly physical experiments. Data-driven evolutionary optimization can be used to search for a set of nondominated tradeoff solutions, where the expensive objective functions are approximated as a surrogate model. In this article, we propose a framework for implementing batched data-driven evolutionary multiobjective optimization (EMO). It is so general that any off-the-shelf EMO algorithms can be applied in a plug-in manner. There are two unique components: 1) based on the Karush–Kuhn–Tucker conditions, a manifold interpolation approach that explores more diversified solutions with a convergence guarantee along the manifold of the approximated Pareto-optimal set and 2) a batch recommendation approach that reduces the computational time of the data-driven evolutionary optimization process by evaluating multiple samples at a time in parallel. Comparing against seven state-of-the-art surrogate-assisted evolutionary algorithms, experiments on 168 benchmark test problem instances with various properties and a real-world application on hyper-parameter optimization fully demonstrate the effectiveness and superiority of our proposed framework, which is featured with a faster convergence and a stronger resilience to various Pareto-optimal front shapes.
@article{LiC23,
author = {Ke Li and
Renzhi Chen},
title = {Batched Data-Driven Evolutionary Multiobjective Optimization Based on Manifold Interpolation},
journal = {IEEE Transactions on Evolutionary Computation},
year = {2023},
volume = {27},
number = {1},
pages = {126--140},
url = {https://ieeexplore.ieee.org/document/9744035},
doi = {10.1109/TEVC.2022.3162993}
}
-
MTLP-JR: Multi-Task Learning-Based Prediction for Joint Ranking in Neural Architecture Search
Bo Lyu, Longfei Lu, Maher Hamdi, Shiping Wen, Yin Yang, Ke Li
Computers & Electrical Engineering, 105: 108474, 2023
10.1016/J.COMPELECENG.2022.108474
Abs | PDF | BiB | Cited by 4
At present, great attentions have been paid to multi-objective neural architecture search (NAS) and resource-aware NAS for their comprehensive consideration of the overall evaluation of architectures, including inference latency, precision, and model scale. However NAS also exacerbates the ever-increasing cost (engineering, time complexity, computation resource). Aiming to alleviate this, the reproducible NAS research releases the benchmark, which includes the metrics (e.g. Accuracy, Latency, and Parameters) of representative models from the typical search space on specific tasks. Motivated by the multi-objective NAS, resource-aware NAS, and reproducible NAS, this paper dedicates to binary-relation prediction (Latency, Accuracy), which is a more reasonable and effective way to satisfy the general NAS scenarios with less cost. We conduct a reproducible NAS study on the MobileNet-based search space and release the dataset. Further, we first propose the modeling of common features among prediction tasks (Latency, Accuracy, Parameters, and FLOPs), which will facilitate the prediction of individual tasks, and creatively formulate the architecture ranking prediction with a multi-task learning framework. Eventually, the proposed multi-task learning based binary-relation prediction model reaches the performance of 94.3% on Latency and 85.02% on Top1 Accuracy even with only 100 training points, which outperforms the single-task learning based model.
@article{LyuLHWYL23,
author = {Bo Lyu and
Longfei Lu and
Maher Hamdi and
Shiping Wen and
Yin Yang and
Ke Li},
title = {MTLP-JR: Multi-task learning-based prediction for joint ranking in neural architecture search},
journal = {Computers and Electrical Engineering},
year = {2023},
volume = {105},
pages = {108474},
url = {https://www.sciencedirect.com/science/article/pii/S0045790622006899?dgcid=coauthor},
doi = {10.1016/J.COMPELECENG.2022.108474}
}
-
Multioutput Surrogate Assisted Evolutionary Algorithm for Expensive Multi-Modal Optimization Problems
Renzhi Chen, Ke Li
Proc. of the 2023 IEEE International Conference on Systems, Man, and Cybernetics (SMC'23)
10.1109/SMC53992.2023.10394405
Abs | PDF | BiB | Cited by 0
Real-world optimization problems are often computationally expensive and feature multi-modal objective functions. Surrogate-assisted evolutionary optimization has proven to be an effective approach for addressing expensive black-box optimization challenges, but the technique has not been adequately studied in multi-modal situations. In this paper, we propose a simple but effective multi-output surrogate-based approach for empowering surrogate-assisted evolutionary optimization to address expensive multi-modal optimization problems. Specifically, our proposed approach employs a multi-output Gaussian process to capture correlations between data collected from different local areas. Experiments on synthetic benchmark test problems demonstrate the effectiveness of our proposed algorithm against five state-of-the-art peer algorithms.
@inproceedings{Chen23,
author = {Renzhi Chen and
Ke Li},
title = {Multioutput Surrogate Assisted Evolutionary Algorithm for Expensive Multi-Modal Optimization Problems},
booktitle = {2023 IEEE International Conference on Systems, Man, and Cybernetics (SMC)},
year = {2023},
volume = {11},
pages = {5176--5181},
url = {https://ieeexplore.ieee.org/document/10394405},
doi = {10.1109/SMC53992.2023.10394405}
}
-
Empirical Studies of Resampling Strategies in Noisy Evolutionary Multi-Objective Optimization
Shasha Zhou, Ke Li
Proc. of the 2023 IEEE International Conference on Systems, Man, and Cybernetics (SMC'23)
10.1109/SMC53992.2023.10394174
Abs | PDF | BiB | Cited by 0
Optimization problems are ubiquitous in real-world engineering scenarios where the goals are to enhance interested aspects such as efficiency, productivity, and profitability. However, solving practical optimization problems could be non-trivial, partly due to the presence of a wide range of noises, including environmental noises, model biases, time-domain variations, measurement uncertainties and many other uncontrolled variables. In this paper, we empirically study the effect of noise range, sample size and resampling type on the solution quality of MOEAs when noise is added to decision variables. Our empirical results, conducted on three commonly used Multi-Objective Optimization Problems (MOEAs), i.e. NSGA-II, MOEA/D and IBEA, demonstrate that noise range has more significant impact on the robustness of optimization algorithms compared to sample size and resampling type. In addition, we introduce the concept of bad point, which is able to illustrate how noise affects the performance of different MOEAs.
@inproceedings{ZhouL23,
author = {Shasha Zhou and
Ke Li},
title = {Empirical Studies of Resampling Strategies in Noisy Evolutionary Multi-Objective Optimization},
booktitle = {2023 IEEE International Conference on Systems, Man, and Cybernetics (SMC)},
year = {2023},
pages = {1715--1720},
url = {https://ieeexplore.ieee.org/document/10394174},
doi = {10.1109/SMC53992.2023.10394174}
}
-
Preference-Based Multi-Objective Optimization with Gaussian Process
Tian Huang, Ke Li
Proc. of the 2023 IEEE International Conference on Systems, Man, and Cybernetics (SMC'23)
PDF | BiB
-
A Multi-Population Evolutionary Algorithm Using New Cooperative Mechanism for Solving Multi-Objective Problems with Multi-Constraint
Juan Zou, Ruiqing Sun, Yuan Liu, Yaru Hu, Shengxiang Yang, Jinhua Zheng, Ke Li
IEEE Trans. Evolutionary Computation (TEVC)
10.1109/TEVC.2023.3260306
Abs | PDF | BiB | Cited by 84
In science and engineering, multiobjective optimization problems (MOPs) usually contain multiple complex constraints, which poses a significant challenge in obtaining the optimal solution. This article aims to solve the challenges brought by multiple complex constraints. First, this article analyzes the relationship between single-constrained Pareto front (SCPF) and their common Pareto front (PF) subconstrained PF (SubCPF). Next, we discussed the SCPF, SubCPF, and unconstraint PF (UPF)’s help to solve constraining PF (CPF). Then, further discusses what kind of cooperation should be used between multiple populations constrained multiobjective optimization algorithm (CMOEA) to better deal with multiconstrained MOPs (mCMOPs). At the same time, based on the discussion in this article, we propose a new multipopulation CMOEA called MCCMO, which uses a new cooperation mechanism. MCCMO uses C+2 (C is the number of constraints) populations to find the UPF, SCPF, and SubCPF at an appropriate time. Furthermore, MCCMO uses the newly proposed activation dormancy detection (ADD) to accelerate the optimization process and uses the proposed combine occasion detection (COD) to find the appropriate time to find the SubCPF. The performance on 32 mCMOPs and real-world mCMOPs shows that our algorithm can obtain competitive solutions on MOPs with multiple constraints.
@article{Zou2024,
author = {Juan Zou and
Ruiqing Sun and
Yuan Liu and
Yaru Hu and
Shengxiang Yang and
Jinhua Zheng and
Ke Li},
title = {A Multipopulation Evolutionary Algorithm Using New Cooperative Mechanism for Solving Multiobjective Problems With Multiconstraint},
journal = {IEEE Transactions on Evolutionary Computation},
year = {2024},
volume = {28},
number = {1},
pages = {267--280},
url = {https://ieeexplore.ieee.org/abstract/document/10078268},
doi = {10.1109/TEVC.2023.3260306}
}
-
“Why Not Looking backward?” A Robust Two-Step Method to Automatically Terminate Bayesian Optimization
Shuang Li, Ke Li, Wei Li
Proc. of the 37th Conference on Neural Information Processing Systems (NeurIPS'23)
Abs | PDF | Code | BiB | ≈ 26.1%
Bayesian Optimization (BO) is a powerful method for tackling expensive black-box optimization problems. As a sequential model-based optimization strategy, BO iteratively explores promising solutions until a predetermined budget, either iterations or time, is exhausted. The decision on when to terminate BO significantly influences both the quality of solutions and its computational efficiency. In this paper, we propose a simple, yet theoretically grounded, two-step method for automatically terminating BO. Our core concept is to proactively identify if the search is within a convex region by examining previously observed samples. BO is halted once the local regret within this convex region falls below a predetermined threshold. To enhance numerical stability, we propose an approximation method for calculating the termination indicator by solving a bilevel optimization problem. We conduct extensive empirical studies on diverse benchmark problems, including synthetic functions, reinforcement learning, and hyperparameter optimization. Experimental results demonstrate that our proposed method saves up to approximately 80% computational budget yet is with an order of magnitude smaller performance degradation, comparing against the other peer methods. In addition, our proposed termination method is robust in terms of the setting of its termination criterion.
@inproceedings{LiLL23,
author = {Shuang Li and
Ke Li and
Wei Li},
title = {“Why Not Looking backward?” A Robust Two-Step Method to Automatically Terminate Bayesian Optimization},
booktitle = {NeurIPS'23: Proc. of 37th Conference on Neural Information Processing Systems},
pages = {1--12},
year = {2023},
url = {https://openreview.net/pdf?id=IMiGRqltQQ}
}
-
CamoPatch: An Evolutionary Strategy for Generating Camoflauged Adversarial Patches
Phoenix Williams, Ke Li
Proc. of the 37th Conference on Neural Information Processing Systems (NeurIPS'23)
PDF | BiB | ≈ 26.1%
-
Evolutionary Multi-objective Instruction Optimization via Large Language Model-based Instruction Operators
Heng Yang, Ke Li
Proc. of the 2023 Conference on Empirical Methods in Natural Language Processing (EMNLP'23)
Abs | PDF | BiB
Instruction-based language modeling has received significant attention in pretrained language models. However, the efficiency of instruction engineering remains low and hinders the development of instruction studies. Recent studies have focused on automating instruction generation, but they primarily aim to improve performance without considering other crucial objectives that impact instruction quality, such as instruction length and perplexity. Therefore, we propose a novel approach (i.e., InstOptima) that treats instruction generation as an evolutionary multi-objective optimization problem. In contrast to text edition-based methods, our approach utilizes a large language model (LLM) to simulate instruction operators, including mutation and crossover. Furthermore, we introduce an objective-guided mechanism for these operators, allowing the LLM to comprehend the objectives and enhance the quality of the generated instructions. Experimental results demonstrate improved fine-tuning performance and the generation of a diverse set of high-quality instructions.
@inproceedings{YangL23a,
author = {Heng Yang and
Ke Li},
editor = {Houda Bouamor and
Juan Pino and
Kalika Bali},
title = {InstOptima: Evolutionary Multi-objective Instruction Optimization
via Large Language Model-based Instruction Operators},
booktitle = {EMNLP'23: Findings of the Association for Computational Linguistics: {EMNLP} 2023},
pages = {13593--13602},
publisher = {Association for Computational Linguistics},
year = {2023},
url = {https://aclanthology.org/2023.findings-emnlp.907},
timestamp = {Wed, 13 Dec 2023 17:20:20 +0100},
biburl = {https://dblp.org/rec/conf/emnlp/0008L23.bib},
bibsource = {dblp computer science bibliography, https://dblp.org}
}
-
PyABSA: A Modularized Framework for Reproducible Aspect-based Sentiment Analysis
Heng Yang, Ke Li
Proc. of the 32nd ACM International Conference on Information and Knowledge Management (CIKM'23), p. 5117–5122, October, 2023.
10.1145/3583780.3614752
Abs | PDF | Code | BiB | ≈ 35% | Cited by 45
The advancement of aspect-based sentiment analysis (ABSA) has highlighted the lack of a user-friendly framework that can significantly reduce the difficulty of reproducing state-of-the-art ABSA performance, especially for beginners. To meet this demand, we present PyABSA, a modularized framework built on PyTorch for reproducible ABSA. To facilitate ABSA research, PyABSA supports several ABSA subtasks, including aspect term extraction, aspect sentiment classification, and end-to-end aspect-based sentiment analysis. With just a few lines of code, the result of a model on a specific dataset can be reproduced. With a modularized design, PyABSA can also be flexibly extended to incorporate new models, datasets, and other related tasks. Additionally, PyABSA highlights its data augmentation and annotation features, which significantly address data scarcity. The project is available at: https://github.com/yangheng95/PyABSA.
@inproceedings{YangZL23,
author = {Heng Yang and
Chen Zhang and
Ke Li},
title = {PyABSA: A Modularized Framework for Reproducible Aspect-based Sentiment Analysis},
booktitle = {Proceedings of the 32nd ACM International Conference on Information and Knowledge Management},
year = {2023},
pages = {5117--5122},
url = {https://dl.acm.org/doi/10.1145/3583780.3614752},
doi = {10.1145/3583780.3614752}
}
-
A Surrogate Assisted Evolutionary Strategy for Image Approximation by Density-Ratio Estimation
Phoenix Williams, Ke Li
Proc. of 2023 IEEE Congress on Evolutionary Computation (CEC'23), p. 1–8, July, 2023.
10.1109/CEC53210.2023.10254060
Abs | PDF | BiB | Cited by 1
The use of evolutionary strategies in generating images is a common practice within the computational art community. In particular, a popular approach is to approximate a target image using overlapping, semi-transparent shapes and optimizing their attributes to increase similarity to the target image. However, existing methods usually require millions of fitness evaluations to construct good approximations. Within the evolutionary computation and machine learning communities, the use of surrogates has shown to decrease the number of fitness evaluations while achieving state-of-the-art results. Despite the gained traction of surrogate-assisted algorithms, their use within the computational art community is nonexistent. To address this, we extend the previous work of Bayesian Optimization by density-ratio estimation (BORE) to the image approximation task. By estimating the probability of improvement acquisition function using a convolutional probabilistic classifier, we search for solutions that maximize the acquisition function using an evolutionary strategy. By conducting experiments on six different styled target images, we demonstrate the superior performance achieved with the use of surrogate assistance.
@inproceedings{WilliamsL23,
author = {Phoenix Williams and
Ke Li and
Geyong Min},
title = {A Surrogate Assisted Evolutionary Strategy for Image Approximation by Density-Ratio Estimation},
booktitle = {2023 IEEE Congress on Evolutionary Computation (CEC)},
year = {2023},
pages = {1--8},
url = {https://doi.org/10.1109/CEC53210.2023.10254060},
doi = {10.1109/CEC53210.2023.10254060}
}
-
Exploring Structural Similarity in Fitness Landscapes via Graph Data Mining: A Case Study on Number Partitioning Problems
Mingyu Huang, Ke Li
Porc. of the 32nd International Joint Conference on Artificial Intelligence (IJCAI'23), p. 5595–5603, August, 2023.
10.24963/ijcai.2023/621
Abs | PDF | Supp | BiB | ≈ 15% | Cited by 3
One of the most common problem-solving heuristics is by analogy. For a given problem, a solver can be viewed as a strategic walk on its fitness landscape. Thus if a solver works for one problem instance, we expect it will also be effective for other instances whose fitness landscapes essentially share structural similarities with each other. However, due to the black-box nature of combinatorial optimization, it is far from trivial to infer such similarity in real-world scenarios. To bridge this gap, by using local optima network as a proxy of fitness landscapes, this paper proposed to leverage graph data mining techniques to conduct qualitative and quantitative analyses to explore the latent topological structural information embedded in those landscapes. In our experiments, we use the number partitioning problem as the case and our empirical results are inspiring to support the overall assumption of the existence of structural similarity between landscapes within neighboring dimensions. Besides, experiments on simulated annealing demonstrate that the performance of a meta-heuristic solver is similar on structurally similar landscapes.
@inproceedings{HuangL23,
author = {Mingyu Huang and
Ke Li},
title = {Exploring Structural Similarity in Fitness Landscapes via Graph Data Mining: A Case Study on Number Partitioning Problems},
year = {2023},
pages = {5595--5603},
url = {https://doi.org/10.24963/ijcai.2023/621},
doi = {10.24963/ijcai.2023/621}
}
-
Boosting Text Augmentation via Hybrid Instance Filtering Framework
Heng Yang, Ke Li
Findings of the 61st Annual Meeting of the Association for Computational Linguistics (ACL'23), p. 1652–1669, July, 2023.
10.18653/v1/2023.findings-acl.105
Abs | PDF | BiB | ≈ 40.6% | Cited by 5
Text augmentation is an effective technique for addressing the problem of insufficient data in natural language processing. However, existing text augmentation methods tend to focus on few-shot scenarios and usually perform poorly on large public datasets. Our research indicates that existing augmentation methods often generate instances with shifted feature spaces, which leads to a drop in performance on the augmented data (for example, EDA generally loses approximately 2% in aspect-based sentiment classification). To address this problem, we propose a hybrid instance-filtering framework (BoostAug) based on pre-trained language models that can maintain a similar feature space with natural datasets. BoostAug is transferable to existing text augmentation methods (such as synonym substitution and back translation) and significantly improves the augmentation performance by 2-3% in classification accuracy. Our experimental results on three classification tasks and nine public datasets show that BoostAug addresses the performance drop problem and outperforms state-of-the-art text augmentation methods. Additionally, we release the code to help improve existing augmentation methods on large datasets.
@inproceedings{YangL23b,
author = {Heng Yang and
Ke Li},
title = {Boosting Text Augmentation via Hybrid Instance Filtering Framework},
booktitle = {Findings of the Association for Computational Linguistics: ACL 2023},
year = {2023},
pages = {1652--1669},
url = {https://doi.org/10.18653/v1/2023.findings-acl.105},
doi = {10.18653/v1/2023.findings-acl.105}
}
-
Single Application Service Deployment in the Edge Environment Based on the E-CARGO Model
Senyue Zhang, Ling Xue, Weiliang Huang, Lu Zhao, Ke Li
Proc. of 26th International Conference on Computer Supported Cooperative Work in Design (CSCWD'23), May, 2023.
10.1109/CSCWD57460.2023.10152752
Abs | PDF | BiB | Cited by 3
The popularization and application of 5G technology is about to open the era of global "data explosion". The traditional cloud computing model shows insufficient service support for resource-sensitive applications, especially in terms of latency, and edge computing can solve this problem by providing service support close to the user request side. This article formalizes the single application service deployment problem (SASDP) using the E-CARGO (Environment-Class, Agent, Role, Group, and Object) model. Through group role assignment (GRA), a high-satisfaction service deployment scheme for single application service deployment is designed, and satisfaction evaluation is established through delay to provide providers with satisfactory deployment solutions and achieve economic benefits. Finally, large-scale simulation experiments are carried out based on Python PuLP platform, and experiments show that our method is better than the baseline method in terms of overall satisfaction, user coverage and economy.
@inproceedings{ZhangXHZL23,
author = {Senyue Zhang and
Ling Xue and
Weiliang Huang and
Lu Zhao and
Ke Li},
title = {Single Application Service Deployment in the Edge Environment Based on the E-CARGO Model},
booktitle = {2023 26th International Conference on Computer Supported Cooperative Work in Design (CSCWD)},
year = {2023},
pages = {1656--1661},
url = {https://ieeexplore.ieee.org/document/10152752},
doi = {10.1109/CSCWD57460.2023.10152752}
}
-
Black-Box Sparse Adversarial Attack via Multi-Objective Optimisation
Phoenix Williams, Ke Li
Proc. of 2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR'23), p. 12291–12301, June, 2023.
10.1109/CVPR52729.2023.01183
Abs | PDF | BiB | ≈ 25% | Cited by 29
Deep neural networks (DNNs) are susceptible to adversarial images, raising concerns about their reliability in safety-critical tasks. Sparse adversarial attacks, which limit the number of modified pixels, have shown to be highly effective in causing DNNs to misclassify. However, existing methods often struggle to simultaneously minimize the number of modified pixels and the size of the modifications, often requiring a large number of queries and assuming unrestricted access to the targeted DNN. In contrast, other methods that limit the number of modified pixels often permit unbounded modifications, making them easily detectable. To address these limitations, we propose a novel multi-objective sparse attack algorithm that efficiently minimizes the number of modified pixels and their size during the attack process. Our algorithm draws inspiration from evolutionary computation and incorporates a mechanism for prioritizing objectives that aligns with an attacker's goals. Our approach outperforms existing sparse attacks on CIFAR-10 and ImageNet trained DNN classifiers while requiring only a small query budget, attaining competitive attack success rates while perturbing fewer pixels. Overall, our proposed attack algorithm provides a solution to the limitations of current sparse attack methods by jointly minimizing the number of modified pixels and their size. Our results demonstrate the effectiveness of our approach in restricted scenarios, highlighting its potential to enhance DNN security.
@inproceedings{WilliamsL23,
author = {Phoenix Neale Williams and
Ke Li},
title = {Black-Box Sparse Adversarial Attack via Multi-Objective Optimisation CVPR Proceedings},
booktitle = {2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
year = {2023},
pages = {12291--12301},
url = {https://doi.org/10.1109/CVPR52729.2023.01183},
doi = {10.1109/CVPR52729.2023.01183}
}
-
Data-Driven Evolutionary Multi-Objective Optimization Based on Multiple-Gradient Descent for Disconnected Pareto Fronts
Renzhi Chen, Ke Li
Proc. of the 12th International Conference on Evolutionary Multi-Criterion Optimization (EMO'23), Springer, p. 56-70, March, 2023
10.1007/978-3-031-27250-9_5
Abs | PDF | Supp | BiB | Cited by 2
Data-driven evolutionary multi-objective optimization (EMO) has been recognized as an effective approach for multi-objective optimization problems with expensive objective functions. The current research is mainly developed for problems with a 'regular' triangle-like Pareto-optimal front (PF), whereas the performance can significantly deteriorate when the PF consists of disconnected segments. Furthermore, the offspring reproduction in the current data-driven EMO does not fully leverage the latent information of the surrogate model. Bearing these considerations in mind, this paper proposes a data-driven EMO algorithm based on multiple-gradient descent. By leveraging the regularity information provided by the up-to-date surrogate model, it is able to progressively probe a set of well distributed candidate solutions with a convergence guarantee. In addition, its infill criterion recommends a batch of promising candidate solutions to conduct expensive objective function evaluations. Experiments on 33 benchmark test problem instances with disconnected PFs fully demonstrate the effectiveness of our proposed method against four selected peer algorithms.
@inproceedings{ChenL23,
author = {Renzhi Chen and
Ke Li},
title = {Data-Driven Evolutionary Multi-objective Optimization Based on Multiple-Gradient Descent for Disconnected Pareto Fronts},
booktitle = {Lecture notes in computer science},
year = {2023},
pages = {56--70},
url = {https://link.springer.com/chapter/10.1007/978-3-031-27250-9_5},
doi = {10.1007/978-3-031-27250-9_5}
}
-
Sparse Adversarial Attack via Bi-Objective Optimization
Phoenix Williams, Ke Li, Geyong Min
Proc. of the 12th International Conference on Evolutionary Multi-Criterion Optimization (EMO'23), Springer, p. 118-133, March, 2023
10.1007/978-3-031-27250-9_9
Abs | PDF | BiB | Cited by 4
Deep neural networks (DNNs) are susceptible to adversarial images, raising concerns about their reliability in safety-critical tasks. Sparse adversarial attacks, which limit the number of modified pixels, have shown to be highly effective in causing DNNs to misclassify. However, existing methods often struggle to simultaneously minimize the number of modified pixels and the size of the modifications, often requiring a large number of queries and assuming unrestricted access to the targeted DNN. In contrast, other methods that limit the number of modified pixels often permit unbounded modifications, making them easily detectable. To address these limitations, we propose a novel multi-objective sparse attack algorithm that efficiently minimizes the number of modified pixels and their size during the attack process. Our algorithm draws inspiration from evolutionary computation and incorporates a mechanism for prioritizing objectives that aligns with an attacker's goals. Our approach outperforms existing sparse attacks on CIFAR-10 and ImageNet trained DNN classifiers while requiring only a small query budget, attaining competitive attack success rates while perturbing fewer pixels. Overall, our proposed attack algorithm provides a solution to the limitations of current sparse attack methods by jointly minimizing the number of modified pixels and their size. Our results demonstrate the effectiveness of our approach in restricted scenarios, highlighting its potential to enhance DNN security.
@inproceedings{WilliamsLGM23,
author = {Phoenix Williams and
Ke Li and
Geyong Min},
title = {Sparse Adversarial Attack via Bi-objective Optimization},
booktitle = {Lecture notes in computer science},
year = {2023},
pages = {118--133},
url = {https://link.springer.com/chapter/10.1007/978-3-031-27250-9_9},
doi = {10.1007/978-3-031-27250-9_9}
}
-
Posterior Decision-Making Based on Decomposition-Driven Knee Point Identification
Ke Li, Haifeng Nie, Huiru Gao, Xin Yao
IEEE Trans. Evolutionary Computation (TEVC), 26(6): 1409-1423, 2022
10.1109/TEVC.2021.3116121
Abs | PDF | Supp | Code | BiB | Cited by 12
Knee points, characterized as a small improvement on one objective can lead to a significant degradation on at least one of the other objectives, are attractive to decision makers (DMs) in multicriterion decision making. This article presents a simple and effective knee point identification (KPI) method to help DMs identify solution(s) of interest from a given set of tradeoff solutions thus facilitating posterior decision making. Our basic idea is to sequentially validate whether a solution is a knee point or not by comparing its localized tradeoff utility with others within its neighborhood characterized from a decomposition perspective. In particular, a solution is a knee point if and only if it has the best-localized tradeoff utility among its neighbors. We implement a GPU version that carries out the KPI in a parallel manner. This GPU version reduces the worst-case complexity from quadratic to linear. The performance of our proposed method is compared with five state-of-the-art KPI methods on 134 test problem instances and two real-world engineering design problems. Empirical results demonstrate its outstanding performance especially on problems with many local knee points. We further validate the usefulness of our proposed method for guiding evolutionary multiobjective optimization algorithms to search for knee points on the fly during the evolutionary process.
@article{LiNGY22,
author = {Ke Li and
Haifeng Nie and
Huiru Gao and
Xin Yao},
title = {Posterior Decision Making Based on Decomposition-Driven Knee Point Identification},
journal = {IEEE Transactions on Evolutionary Computation},
year = {2022},
volume = {26},
number = {6},
pages = {1409--1423},
url = {https://ieeexplore.ieee.org/document/9551284},
doi = {10.1109/TEVC.2021.3116121}
}
-
Distributed UAV Swarm Formation and Collision Avoidance Strategies Over Fixed and Switching Topologies
Jia Wu, Chunbo Luo, Yang Luo, Ke Li
IEEE Trans. Cybernetics (TCYB), 52(10): 10969-10979, 2022
10.1109/TCYB.2021.3132587
Abs | PDF | BiB | Cited by 126
This article proposes a controlling framework for multiple unmanned aerial vehicles (UAVs) to integrate the modes of formation flight and swarm deployment over fixed and switching topologies. Formation strategies enable UAVs to enjoy key collective benefits including reduced energy consumption, but the shape of the formation and each UAV's freedom are significantly restrained. Swarm strategies are thus proposed to maximize each UAV's freedom following simple yet powerful rules. This article investigates the integration and switch between these two strategies, considering the deployment environment factors, such as poor network conditions and unknown and often highly mobile obstacles. We design a distributed formation controller to guide multiple UAVs in orderless states to swiftly reach an intended formation. Inspired by starling birds and similar biological creatures, a distributed collision avoidance controller is proposed to avoid unknown and mobile obstacles. We further illustrated the stability of the controllers over both fixed and switching topologies. The experimental results confirm the effectiveness of the framework.
@article{WuLLL22,
author = {Jia Wu and
Chunbo Luo and
Yang Luo and
Ke Li},
title = {Distributed UAV Swarm Formation and Collision Avoidance Strategies Over Fixed and Switching Topologies},
journal = {IEEE Transactions on Cybernetics},
year = {2022},
volume = {52},
number = {10},
pages = {10969--10979},
url = {https://ieeexplore.ieee.org/document/9663025},
doi = {10.1109/TCYB.2021.3132587}
}
-
Transfer Learning Based Parallel Evolutionary Algorithm Framework for Bi-level Optimization
Lei Chen, Hai-Lin Liu, Kay Chen Tan, Ke Li
IEEE Trans. Evolutionary Computation (TEVC), 26(1): 115-129, 2022
10.1109/TEVC.2021.3095313
Abs | PDF | Code | BiB | Cited by 43
Evolutionary algorithms (EAs) have been recognized as a promising approach for bilevel optimization. However, the population-based characteristic of EAs largely influences their efficiency and effectiveness due to the nested structure of the two levels of optimization problems. In this article, we propose a transfer learning-based parallel EA (TLEA) framework for bilevel optimization. In this framework, the task of optimizing a set of lower level problems parameterized by upper level variables is conducted in a parallel manner. In the meanwhile, a transfer learning strategy is developed to improve the effectiveness of each lower level search (LLS) process. In practice, we implement two versions of the TLEA: the first version uses the covariance matrix adaptation evolutionary strategy and the second version uses the differential evolution as the evolutionary operator in lower level optimization. The experimental studies on two sets of widely used bilevel optimization benchmark problems are conducted, and the performance of the two TLEA implementations is compared to that of four well-established evolutionary bilevel optimization algorithms to verify the effectiveness and efficiency of the proposed algorithm framework.
@article{ChenLTL22,
author = {Lei Chen and
Hai-Lin Liu and
Kay Chen Tan and
Ke Li},
title = {Transfer Learning-Based Parallel Evolutionary Algorithm Framework for Bilevel Optimization},
journal = {IEEE Transactions on Evolutionary Computation},
year = {2022},
volume = {26},
number = {1},
pages = {115--129},
url = {https://ieeexplore.ieee.org/abstract/document/9476019},
doi = {10.1109/TEVC.2021.3095313}
}
-
Preference based Multi-Objective Reinforcement Learning for Multi-Microgrid System Optimization Problem in Smart Grid
Jiangjiao Xu, Ke Li, Mohammad Abusara
Memetic Computing (MC), 14(2): 225-235, 2022
10.1007/s12293-022-00357-w
Abs | PDF | BiB | Cited by 28
Abstract Grid-connected microgrids comprising renewable energy, energy storage systems and local load, play a vital role in decreasing the energy consumption of fossil diesel and greenhouse gas emissions. A distribution power network connecting several microgrids can promote more potent and reliable operations to enhance the security and privacy of the power system. However, the operation control for a multi-microgrid system is a big challenge. To design a multi-microgrid power system, an intelligent multi-microgrids energy management method is proposed based on the preference-based multi-objective reinforcement learning (PMORL) techniques. The power system model can be divided into three layers: the consumer layer, the independent system operator layer, and the power grid layer. Each layer intends to maximize its benefit. The PMORL is proposed to lead to a Pareto optimal set for each object to achieve these objectives. A non-dominated solution is decided to execute a balanced plan not to favor any particular participant. The preference-based results show that the proposed method can effectively learn different preferences. The simulation outcomes confirm the performance of the PMORL and verify the viability of the proposed method.
@article{XuLA22,
author = {Jiangjiao Xu and
Ke Li and
Mohammad Abusara},
title = {Preference based multi-objective reinforcement learning for multi-microgrid system optimization problem in smart grid},
journal = {Memetic Computing},
year = {2022},
volume = {14},
number = {2},
pages = {225--235},
url = {https://link.springer.com/article/10.1007/s12293-022-00357-w},
doi = {10.1007/s12293-022-00357-w}
}
-
Surrogate-Assisted Evolutionary Multi-Objective Optimization for Hardware Design Space Exploration
Renzhi Chen, Ke Li
Proc. of the NeurIPS Workshop on Gaussian Processes, Spatiotemporal Modeling, and Decision-making Systems, 2022
Abs | PDF | BiB
Hardware design space exploration (DSE) aims to find a suitable micro-architecture for the dedicated hardware accelerators. It is a computationally expensive black-box optimization problem with more than one conflicting performance indicator. Surrogate-assisted evolutionary algorithm is a promising framework for expensive multi-objective optimization problems given its surrogate modeling for handling expensive objective functions and population-based characteristics that search for a set of trade-off solutions simultaneously. However, most, if not all, existing studies mainly focus ‘regular’ Pareto-optimal fronts (PFs), whereas the PF is typically irregular in hardware DSE. In the meanwhile, the gradient information of the differentiable surrogate model(s) is beneficial to navigate a more effective exploration of the search space, but it is yet fully exploited. This paper proposes a surrogate-assisted evolutionary multi-objective optimization based on multiple-gradient descent (MGD) for hardware DSE. Empirical results on both synthetic problems with irregular PFs and real-world hardware DSE cases fully demonstrate the effectiveness and outstanding performance of our proposed algorithm.
-
Are All Training Data Useful? A Empirical Revisit of Subset Selection in Bayesian Optimization
Peili Mao, Ke Li
Proc. of the NeurIPS Workshop on Gaussian Processes, Spatiotemporal Modeling, and Decision-making Systems, 2022
Abs | PDF | BiB
Bayesian optimization (BO) has been widely recognized as a powerful approach for black-box optimization problems with expensive objective function(s). Gaussian process (GP), which has been widely used for surrogate modeling in BO, is notorious for its cubic computational complexity grows with the increase of the amount of evaluated samples. This can lead to a significantly increased computational time for BO due to its sequential decision-making nature. This paper revisit the simple and effective subset selection methods to pick up a small group of representative data from the entire dataset to carry out the training and inference of GP in the context of BO. Empirical studies demonstrate that subset selection methods not only promote the performance of the vanilla BO but also significantly reduce the computational time for up to 98%.
-
Imputation and Forecasting for Multi-Output Gaussian Process in Smart Grid
Jiangjiao Xu, Ke Li
Proc. of the NeurIPS Workshop on Gaussian Processes, Spatiotemporal Modeling, and Decision-making Systems, 2022
Abs |
PDF | BiB
-
Do We Really Need to Use Constraint Violation in Constrained Evolutionary Multi-Objective Optimization?
Shuang Li, Ke Li, Wei Li
Proc. of the 17the International Conference on Parallel Problem Solving from Nature (PPSN'22), Springer, p. 124-137, September, 2022
10.1007/978-3-031-14721-0_9
PDF | BiB | Cited by 4
@article{Li2022,
author = {Shuang Li and
Ke Li and
Wei Li},
title = {Do We Really Need to Use Constraint Violation in Constrained Evolutionary Multi-objective Optimization?},
journal = {Lecture notes in computer science},
year = {2022},
pages = {124--137},
url = {https://link.springer.com/chapter/10.1007/978-3-031-14721-0_9},
doi = {10.1007/978-3-031-14721-0_9}
}
-
Attention-Based Genetic Algorithm for Adversarial Attack in Natural Language Processing
Shasha Zhou, Ke Li, Geyong Min
Proc. of the 17the International Conference on Parallel Problem Solving from Nature (PPSN'22), Springer, p. 341-355, September, 2022
10.1007/978-3-031-14714-2_24
PDF | BiB | Cited by 6
@article{Zhou2022,
author = {Shasha Zhou and
Ke Li and
Geyong Min},
title = {Attention-Based Genetic Algorithm for Adversarial Attack in Natural Language Processing},
journal = {Lecture notes in computer science},
year = {2022},
pages = {341--355},
url = {https://link.springer.com/chapter/10.1007/978-3-031-14714-2_24},
doi = {10.1007/978-3-031-14714-2_24}
}
-
Black-Box Adversarial Attack via Overlapped Shapes
Phoenix Williams, Ke Li, Geyong Min
Proc. of the 24th Annual Conference on Genetic and Evolutionary Computation (GECCO’22), ACM, p. 467-468, July, 2022
10.1145/3520304.3528934
Abs | PDF | BiB | Cited by 2
Deep neural networks (DNNs) have achieved state-of-the-art performance in many tasks but have shown extreme vulnerabilities to attacks generated by adversarial examples. Many works assume an attacker has total access to the targeted model. A realistic assumption is that an attacker has access to the targeted model only by querying some input and observing its predicted class probabilities. In this paper we propose a concept of applying techniques similar to those used within evolutionary-art to generated adversarial images.
@article{Williams2022,
author = {Phoenix Williams and
Ke Li and
Geyong Min},
title = {Black-box adversarial attack via overlapped shapes},
journal = {Proceedings of the Genetic and Evolutionary Computation Conference Companion},
year = {2022},
pages = {467--468},
url = {https://dl.acm.org/doi/10.1145/3520304.3528934},
doi = {10.1145/3520304.3528934}
}
-
Adversarial Attack Example Generation via Genetic Algorithm: A Preliminary Result
Shasha Zhou, Ke Li, Geyong Min
Proc. of the 24th Annual Conference on Genetic and Evolutionary Computation (GECCO’22), ACM, p. 469-470, July, 2022
10.1145/3520304.3528981
Abs | PDF | BiB | Cited by 2
In recent years, some studies showed that deep neural networks (DNNs) are vulnerable to being attacked by small perturbated examples. To satisfy the lexical, grammatical, and semantic constrain, some works proposed using black-box population-based optimization algorithms to attack neural networks in natural language processing. However, they are inefficient enough because they do not consider the characteristics of the text itself. Also, they are slow to close to the decision boundary. In this paper, we propose a more efficient attention based genetic algorithm adversarial attack method, called AGA. We use attention mechanism to pay more attention to the important tokens and utilize the multi-membered strategy to accelerate the search procedure. The result shows that our attack achieves a higher success rate with less than 136% of the number of queries than the existing methods.
@article{Zhou2022,
author = {Shasha Zhou and
Ke Li and
Geyong Min},
title = {Adversarial example generation via genetic algorithm},
journal = {Proceedings of the Genetic and Evolutionary Computation Conference Companion},
year = {2022},
pages = {469--470},
url = {https://dl.acm.org/doi/10.1145/3520304.3528981},
doi = {10.1145/3520304.3528981}
}
-
Decomposition Multi-Objective Optimisation: Current Developments and Future Opportunities
Ke Li, Qingfu Zhang
Proc. of the 24th Annual Conference on Genetic and Evolutionary Computation (GECCO’22), ACM, p. 469-470, July, 2022
10.1145/3520304.3533629
Abs | PDF | BiB | Cited by 0
tutorial Share on Decomposition multi-objective optimisation: current developments and future opportunities Authors: Ke Li University of Exeter and City University of Hong Kong University of Exeter and City University of Hong KongView Profile , Qingfu Zhang University of Exeter and City University of Hong Kong University of Exeter and City University of Hong KongView Profile Authors Info & Claims GECCO '22: Proceedings of the Genetic and Evolutionary Computation Conference CompanionJuly 2022 Pages 1554–1571https://doi.org/10.1145/3520304.3533629Published:19 July 2022Publication History 0citation33DownloadsMetricsTotal Citations0Total Downloads33Last 12 Months33Last 6 weeks3 Get Citation AlertsNew Citation Alert added!This alert has been successfully added and will be sent to:You will be notified whenever a record that you have chosen has been cited.To manage your alert preferences, click on the button below.Manage my Alerts New Citation Alert!Please log in to your account Save to BinderSave to BinderCreate a New BinderNameCancelCreateExport CitationPublisher SiteGet Access
@article{Li2022,
author = {Ke Li and
Qingfu Zhang},
title = {Decomposition multi-objective optimisation},
journal = {Proceedings of the Genetic and Evolutionary Computation Conference Companion},
year = {2022},
volume = {01},
pages = {1554--1571},
url = {https://dl.acm.org/doi/10.1145/3520304.3533629},
doi = {10.1145/3520304.3533629}
}
-
Knee Point Identification Based on the Geometric Characteristic
Renzhi Chen, Ke Li
Proc. of the 2021 IEEE International Conference on Systems, Man, and Cybernetics (SMC'21), IEEE, p. 764–769, October, 2021
10.1109/SMC52423.2021.9658848
Abs | PDF | BiB | Cited by 3
The ultimate goal of multi-objective optimisation is to help decision makers (DMs) identify solution(s) of interest. However, providing the DMs with a large amount of the trade-off alternatives not only increase their workload, but also add irrelevant noise to the decision-making process. Without any prior knowledge, knee points, characterised as their smallest trade-off loss at all objectives, are attractive to decision makers in multi-criterion decision-making. In this paper, we propose a simple but effective knee point identification method based on Voronoi diagram. It divides the objective space into several Voronoi cells to capture the geometric characteristics of the underlying trade-off solution set. Thereafter, the knee points are identified as those having a local Voronoi distance. Empirical results demonstrate that our proposed method is able to identify knee points located in both convex and concave part of the corresponding Pareto-optimal front.
@inproceedings{Chen021,
author = {Renzhi Chen and
Ke Li},
title = {Knee Point Identification Based on the Geometric Characteristic},
booktitle = {2021 IEEE International Conference on Systems, Man, and Cybernetics (SMC)},
year = {2021},
pages = {764--769},
url = {https://ieeexplore.ieee.org/abstract/document/9658848},
doi = {10.1109/SMC52423.2021.9658848}
}
-
Large-Scale Evolutionary Optimization via Multi-Task Random Grouping
Phoenix Williams, Ke Li
Proc. of the 2021 IEEE International Conference on Systems, Man, and Cybernetics (SMC'21), IEEE, p. 778–783, October, 2021
10.1109/SMC52423.2021.9659276
Abs | PDF | BiB | Cited by 2
Evolutionary Algorithms (EA) are known to suffer from the curse of dimensionality resulting in poor performances when handling large-scale problems. Cooperative coevolution aims to overcome these issues in a divide and conquer approach by decomposing the original problem into several lower-dimensional sub-problems. For each sub-problem a chosen EA is applied for a defined number of function evaluations. This is repeated in a round robin like fashion until a terminating condition is met. A recently proposed area in the evolutionary computation field is the evolutionary multitask optimization (EMTO) framework. By jointly optimising several tasks, EMTO aims to exploit beneficial information across multiple tasks to improve the performance compared to optimising each task in isolation. In this paper, we consider a large-scale problem as a multi-task optimization problem by considering each sub-problem as an independent task. Applying an EMTO algorithm, knowledge transfer across sub-problems is carried out explicitly to improve the optimization of each sub-problem. We evaluate the effectiveness of our proposed algorithms empirically on a suite of separable and non-separable benchmark problems of varying dimensions.
@inproceedings{WilliamsLM21,
author = {Phoenix Neale Williams and
Ke Li and
Geyong Min},
title = {Large-Scale Evolutionary Optimization via Multi-Task Random Grouping},
booktitle = {2021 IEEE International Conference on Systems, Man, and Cybernetics (SMC)},
year = {2021},
pages = {778--783},
url = {https://ieeexplore.ieee.org/abstract/document/9659276},
doi = {10.1109/SMC52423.2021.9659276}
}
-
Transfer Bayesian Optimization for Expensive Black-Box Optimization in Dynamic Environment
Renzhi Chen, Ke Li
Proc. of the 2021 IEEE International Conference on Systems, Man, and Cybernetics (SMC'21), IEEE, p. 1374–1379, October, 2021
10.1109/SMC52423.2021.9659200
Abs | PDF | BiB | Cited by 10
Expensive black-box optimization in dynamic environments is a challenging but important task since many real-world problems are changing over time and are computationally costly. Bayesian optimization has been widely recognized as an effective approach for tackling expensive black-box optimization in a static environment whereas it has rarely been studied for in dynamic environments. This paper proposes a simple but effective method to empower Bayesian optimization to solve dynamic optimization problems. It augments the covariance function with the measurement of the relationship between historical observations and the current ones. By doing so, the Bayesian optimization is able to leverage the observations from the previous time step to jump start the optimization in the new environment with a strictly limited computational budget. Experiments on synthetic benchmark test problems and a real-world case study demonstrate the effectiveness of our proposed algorithm.
@inproceedings{ChenL21,
author = {Renzhi Chen and
Ke Li},
title = {Transfer Bayesian Optimization for Expensive Black-Box Optimization in Dynamic Environment},
booktitle = {2021 IEEE International Conference on Systems, Man, and Cybernetics (SMC)},
year = {2021},
pages = {1374--1379},
url = {https://ieeexplore.ieee.org/abstract/document/9659200},
doi = {10.1109/SMC52423.2021.9659200}
}
-
ADMM-based OPF Problem Against Cyber Attacks in Smart Grid
Jiangjiao Xu, Ke Li, Mohammad Abusara, Yan Zhang
Proc. of the 2021 IEEE International Conference on Systems, Man, and Cybernetics (SMC'21), IEEE, p. 1418–1423, October, 2021
10.1109/SMC52423.2021.9658699
Abs | PDF | BiB | Cited by 2
In the smart grid, the application of information and communication technology (ICT) significantly promotes the efficiency of energy generation and consumption system. However, the integration of intelligence and cyber system to a smart grid can result in serious cyber security concerns and presents the entire power system more vulnerable to be attacked. In this paper, a new cyber attack model with the alternating direction multiplier method (ADMM) based optimal power flow (OPF) problem is introduced and exploited by malicious attackers. To deal with this challenge, a defence mechanism is presented to not only detect the existence of data injection attack, but also mitigate the potential impact to improve the stability of the smart grid. This scheme takes the power measurements between neighbouring nodes into account, and distinguishes the time-delay attack and bad data injection attack by monitoring the measurement variations between received data and predicted data based on the artificial neural network (ANN) algorithm. The simulation results demonstrate that the proposed scheme has the capability of detecting and mitigating the stealthy attacks significantly by using a 33-bus power system. Consequently, it is a significant study in the actual smart grid and can minimise the impact of the cyber attack.
@inproceedings{XuLA021,
author = {Jiangjiao Xu and
Ke Li and
Mohammad Abusara and
Yan Zhang},
title = {ADMM-based OPF Problem Against Cyber Attacks in Smart Grid},
booktitle = {2021 IEEE International Conference on Systems, Man, and Cybernetics (SMC)},
year = {2021},
pages = {1418--1423},
url = {https://ieeexplore.ieee.org/abstract/document/9658699},
doi = {10.1109/SMC52423.2021.9658699}
}
-
An Enhancement of the NSGA-II Reliability Optimization using Extended Kalman Filter Based Initialization
Savas Yuec, Ke Li
Proc. of the 2021 20th UK Workshop on Computational Intelligence (UKCI'21), Springer, p. 121–128, September, 2021
10.1007/978-3-030-87094-2_11
PDF | BiB | Cited by 0
@inproceedings{YuceL21,
author = {Savas Yuce and
Ke Li},
title = {An Enhancement of the NSGA-II Reliability Optimization Using Extended Kalman Filter Based Initialization},
booktitle = {Advances in intelligent systems and computing},
year = {2021},
pages = {121--128},
url = {https://link.springer.com/chapter/10.1007/978-3-030-87094-2_11},
doi = {10.1007/978-3-030-87094-2_11}
}
-
Empirical Studies on the Role of the Decision Maker in Interactive Evolutionary Multi-Objective Optimization
Guiyu Lai, Minhui Liao, Ke Li
Proc. of the 2021 IEEE Congress on Evolutionary Computation (CEC'21), IEEE, p. 1–8, June, 2021
10.1109/CEC45853.2021.9504980
Abs | PDF | BiB | Cited by 9
The interactive evolutionary multi-objective optimization (IEMO) algorithms aim to learn and utilize the preference information from the decision maker (DM) during the optimization process to guide the search towards preferred solutions. In this paper, we are devoted to figuring out the effects of interaction patterns, DM calls, preference changes, and DM inconsistencies on the quality of the solutions generated by the IEMO algorithms. The investigation is done in the context of I-MOEA/D-PLVF algorithm, a recently proposed interactive optimization algorithm based on MOEA/D.The experimental results indicate that different interaction patterns and the number of DM calls do result in significant impacts on the quality of the obtained solutions generated by the IEMO algorithm used in our experiments. Meanwhile, preference changes and DM inconsistencies in the process of interactions will impose irreversibly negative effects on obtained solutions.
@inproceedings{LaiL021,
author = {Guiyu Lai and
Minhui Liao and
Ke Li},
title = {Empirical Studies on the Role of the Decision Maker in Interactive Evolutionary Multi-Objective Optimization},
booktitle = {2021 IEEE Congress on Evolutionary Computation (CEC)},
year = {2021},
pages = {185--192},
url = {https://ieeexplore.ieee.org/document/9504980},
doi = {10.1109/CEC45853.2021.9504980}
}
-
Empirical Study of Correlations in the Fitness Landscapes of Combinatorial Optimization Problems
Longfei Zhang, Ke Li, Shi Gu
Proc. of the 23th Annual Conference on Genetic and Evolutionary Computation (GECCO’21), ACM, p. 247–248, July, 2021.
10.1145/3449726.3459528
Abs | PDF | BiB | Cited by 1
One of the most common problem-solving heuristics is by analogy. For a given problem, a solver can be viewed as a strategic walk on its fitness landscape. Thus if a solver works for one problem, it is anticipated that it will be effective for problems within the same category whose fitness landscapes essentially share structural similarity with each other. However, due to the black-box nature, it is far from trivial to infer such similarity in real-world scenarios. To bridge this gap, this paper proposes two alternative approaches to empirically investigate the potential existence of structural similarity among different fitness landscapes. Specifically, we pick up three classic combinatorial optimization problems to constitute the benchmark set. We apply a local optima network construction routine to build a coarse-grained model to represent the fitness landscapes of different problems at various dimensions. Thereafter, we apply a graph embedding method, to empirically investigate the potential existence of correlations with respect to different local optima networks. From our empirical results, we are exciting to find some evidence of the existence of similarity not only for a given problem with various dimensions but also across different problems.
@inproceedings{ZhangLG21,
author = {Longfei Zhang and
Ke Li and
Shi Gu},
title = {Empirical study of correlations in the fitness landscapes of combinatorial optimization problems},
booktitle = {Proceedings of the Genetic and Evolutionary Computation Conference Companion},
year = {2021},
pages = {247--248},
url = {https://dl.acm.org/doi/10.1145/3449726.3459528},
doi = {10.1145/3449726.3459528}
}
-
An Improved Two-Archive Evolutionary Algorithm for Constrained Multi-Objective Optimization
Xinyu Shan, Ke Li
Proc. of the 11th International Conference on Evolutionary Multi-Criterion Optimization (EMO'21), Springer LNCS, volume 12654, p. 235–247, March, 2021
10.1007/978-3-030-72062-9_19
PDF | BiB | Cited by 5
@inproceedings{ShanL21,
author = {Xinyu Shan and
Ke Li},
title = {An Improved Two-Archive Evolutionary Algorithm for Constrained Multi-objective Optimization},
booktitle = {Lecture notes in computer science},
year = {2021},
pages = {235--247},
url = {https://doi.org/10.1007/978-3-030-72062-9_19},
doi = {10.1007/978-3-030-72062-9_19}
}
-
Multi-Objective Reinforcement Learning based Multi-Microgrid System Optimisation Problem
Jiangjiao Xu, Ke Li, Mohammad Abusara
Proc. of the 11th International Conference on Evolutionary Multi-Criterion Optimization (EMO'21), Springer LNCS, volume 12654, p. 684–696, March, 2021
10.1007/978-3-030-72062-9_54
PDF | BiB | Cited by 4
@inproceedings{XuLA21,
author = {Jiangjiao Xu and
Ke Li and
Mohammad Abusara},
title = {Multi-objective Reinforcement Learning Based Multi-microgrid System Optimisation Problem},
booktitle = {Lecture notes in computer science},
year = {2021},
pages = {684--696},
url = {https://doi.org/10.1007/978-3-030-72062-9_54},
doi = {10.1007/978-3-030-72062-9_54}
}
-
Parallel Algorithms for Multiobjective Virtual Network Function Placement Problem
Joseph Billingsley, Ke Li, Wang Miao, Geyong Min, Nektarios Georgalas
Proc. of the 11th International Conference on Evolutionary Multi-Criterion Optimization (EMO'21), Springer LNCS, volume 12654, p. 708–720, March, 2021
10.1007/978-3-030-72062-9_56
PDF | BiB | Cited by 2
@inproceedings{BillingsleyLMMG21,
author = {Joseph Billingsley and
Ke Li and
Wang Miao and
Geyong Min and
Nektarios Georgalas},
title = {Parallel Algorithms for the Multiobjective Virtual Network Function Placement Problem},
booktitle = {Lecture notes in computer science},
year = {2021},
pages = {708--720},
url = {https://doi.org/10.1007/978-3-030-72062-9_56},
doi = {10.1007/978-3-030-72062-9_56}
}
-
Does Preference Always Help? A Holistic Study on Preference-Based Evolutionary Multi-Objective Optimisation Using Reference Points
Ke Li, Minhui Liao, Kalyanmoy Deb, Geyong Min, Xin Yao
IEEE Trans. Evolutionary Computation (TEVC), 24(6): 1078–1096, 2020.
10.1109/TEVC.2020.2987559
Abs | PDF | Code | BiB | Cited by 63
The ultimate goal of multiobjective optimization is to help a decision maker (DM) identify solution(s) of interest (SOI) achieving satisfactory tradeoffs among multiple conflicting criteria. This can be realized by leveraging DM's preference information in evolutionary multiobjective optimization (EMO). No consensus has been reached on the effectiveness brought by incorporating preference in EMO (either a priori or interactively) versus a posteriori decision making after a complete run of an EMO algorithm. Bearing this consideration in mind, this article: 1) provides a pragmatic overview of the existing developments of preference-based EMO (PBEMO) and 2) conducts a series of experiments to investigate the effectiveness brought by preference incorporation in EMO for approximating various SOI. In particular, the DM's preference information is elicited as a reference point, which represents her/his aspirations for different objectives. The experimental results demonstrate that preference incorporation in EMO does not always lead to a desirable approximation of SOI if the DM's preference information is not well utilized, nor does the DM elicit invalid preference information, which is not uncommon when encountering a black-box system. To a certain extent, this issue can be remedied through an interactive preference elicitation. Last but not the least, we find that a PBEMO algorithm is able to be generalized to approximate the whole PF given an appropriate setup of preference information.
@article{LiLDMY20,
author = {Ke Li and
Minhui Liao and
Kalyanmoy Deb and
Geyong Min and
Xin Yao},
title = {Does Preference Always Help? A Holistic Study on Preference-Based Evolutionary Multiobjective Optimization Using Reference Points},
journal = {IEEE Transactions on Evolutionary Computation},
year = {2020},
volume = {24},
number = {6},
pages = {1078--1096},
url = {https://ieeexplore.ieee.org/document/9066927},
doi = {10.1109/TEVC.2020.2987559}
}
-
Reference Point Based Multi-Objective Optimization of Reservoir Operation: a Comparison of Three Algorithms
Rong Tang, Ke Li, Wei Ding, Yuntao Wang, Huicheng Zhou, Guangtao Fu
Water Resources Management, 34: 1005–1020, 2020.
10.1007/s11269-020-02485-9
PDF | BiB | Cited by 27
@article{tangLDWZF20,
author = {Rong Tang and
Ke Li and
Wei Ding and
Yuntao Wang and
Huicheng Zhou and
Guangtao Fu},
title = {Reference Point Based Multi-Objective Optimization of Reservoir Operation: a Comparison of Three Algorithms},
journal = {Water Resources Management},
year = {2020},
volume = {34},
number = {3},
pages = {1005--1020},
url = {http://link.springer.com/article/10.1007/s11269-020-02485-9},
doi = {10.1007/s11269-020-02485-9}
}
-
Evolutionary Many-Objective Optimization Based on Adversarial Decomposition
Mengyuan Wu, Ke Li, Sam Kwong, Qingfu Zhang
IEEE Trans. Cybernetics (TCYB), 50(2): 753–764, 2020.
10.1109/TCYB.2018.2872803
Abs | PDF | Supp | Code | BiB | Cited by 129
The decomposition-based evolutionary algorithm has become an increasingly popular choice for posterior multiobjective optimization. Facing the challenges of an increasing number of objectives, many techniques have been developed which help to balance the convergence and diversity. Nevertheless, according to a recent study by Ishibuchi et al., due to the predefined search directions toward the ideal point, their performance strongly depends on the Pareto front (PF) shapes, especially the orientation of the PFs. To balance the convergence and diversity for decomposition-based methods and to alleviate their performance dependence on the orientation of the PFs, this paper develops an adversarial decomposition method for many-objective optimization, which leverages the complementary characteristics of different subproblem formulations within a single paradigm. More specifically, two populations are co-evolved by two subproblem formulations with different contours and adversarial search directions. To avoid allocating redundant computational resources to the same region of the PF, the two populations are matched into one-to-one solution pairs according to their working regions upon the PF. Each solution pair can at most contribute one principal mating parent during the mating selection process. When comparing nine state-of-the-art many-objective optimizers, we have witnessed the competitive performance of our proposed algorithm on 130 many-objective test problems with various characteristics, including regular and inverted PFs.
@article{WuLKZ20,
author = {Mengyuan Wu and
Ke Li and
Sam Kwong and
Qingfu Zhang},
title = {Evolutionary Many-Objective Optimization Based on Adversarial Decomposition},
journal = {IEEE Transactions on Cybernetics},
year = {2020},
volume = {50},
number = {2},
pages = {753--764},
url = {https://ieeexplore.ieee.org/document/8500750},
doi = {10.1109/TCYB.2018.2872803}
}
-
Performance Analysis of SDN and NFV enabled Mobile Cloud Computing
Joseph Billingsley, Ke Li, Wang Miao, Geyong Min, Nektarios Georgalas
Proc. of the 2020 IEEE Global Communications Conference (GLOBECOM'20), IEEE Press: p. 1–6, December, 2020
10.1109/GLOBECOM42002.2020.9322530
Abs | PDF | BiB | Cited by 6
Mobile Cloud Computing (MCC) is regarded as a promising method to increase the data storage and enhance the processing power of mobile devices. Technologies such as Software Defined Networking (SDN) and Network Function Virtualisation (NFV) will be deployed in MCC to simplify the network management and accelerate mobile service deployment. In order to achieve a deeper understanding of future MCC, we developed a comprehensive analytical model to investigate the performance of MCC in the presence of both NFV service chains and SDN networks. The model is capable of capturing the interactions between SDN and NFV when they share the same underlying physical infrastructure. The end-to-end latency is derived for different scales of service deployments and network configurations. Comprehensive simulation experiments are conducted and the results demonstrate that the proposed analytical model corresponds well with the simulation experiments. In addition, we show how the analytical model can be a useful tool to investigate the impact of centralised SDN control on the performance of NFV traffic transmission.
@inproceedings{BillingsleyMLMG20,
author = {Joseph Billingsley and
Wang Miao and
Ke Li and
Geyong Min and
Nektarios Georgalas},
title = {Performance Analysis of SDN and NFV enabled Mobile Cloud Computing},
booktitle = {GLOBECOM 2020 - 2020 IEEE Global Communications Conference},
year = {2020},
volume = {1},
pages = {1--6},
url = {https://ieeexplore.ieee.org/abstract/document/9322530},
doi = {10.1109/GLOBECOM42002.2020.9322530}
}
-
Knee Point Identification Based on Voronoi Diagram
Haifeng Nie, Huiru Gao, Ke Li
Proc. of the 2020 IEEE Conference on Systems, Man and Cybernetics (SMC'20), IEEE Press: p. 1–6, December, 2020
10.1109/SMC42975.2020.9283262
Abs | PDF | BiB | Cited by 5
Finding preferred solutions is important for DMs to take the next step in solving Multi-objective optimisation problems (MOPs). When no specific preferences are available, knee point(s) are typically considered to be the most preferred solutions in multi-criterion decision-making since their smallest trade-off loss at all objectives. Knee point(s), including concave, convex and edge knee point(s), can reflect some geometry characteristics of the given non-dominated solutions because of its unique location. However, most of contemporary research for knee point identification (KPI) is only designed for convex knee point(s). Based on Voronoi diagram which can effectively reflect the distribution of the given set, we propose a method to identify all three types of knee points from a geometry view. In order to validate our method, we compare the performance of our method with other three state of the art approaches on benchmark problems for knee point identification. Experimental results fully show the effectiveness and competitiveness of our proposed KPI method based on Voronoi diagram (KPIVD) for identifying three types of knee points.
@inproceedings{NieGL20,
author = {Haifeng Nie and
Huiru Gao and
Ke Li},
title = {Knee Point Identification Based on Voronoi Diagram},
booktitle = {2020 IEEE International Conference on Systems, Man, and Cybernetics (SMC)},
year = {2020},
pages = {1081--1086},
url = {https://ieeexplore.ieee.org/document/9283262},
doi = {10.1109/SMC42975.2020.9283262}
}
-
BiLO-CPDP: Bi-Level Programming for Automated Model Discovery in Cross-Project Defect Prediction
Ke Li, Zilin Xiang, Tao Chen, Kay Chen Tan
Proc. of the 35th IEEE/ACM International Conference on Automated Software Engineering (ASE'20), IEEE Press: September, 2020
10.1145/3324884.3416617
Abs | PDF | Code | BiB | Cited by 25
Cross-Project Defect Prediction (CPDP), which borrows data from similar projects by combining a transfer learner with a classifier, have emerged as a promising way to predict software defects when the available data about the target project is insufficient. However, developing such a model is challenge because it is difficult to determine the right combination of transfer learner and classifier along with their optimal hyper-parameter settings. In this paper, we propose a tool, dubbed BiLO-CPDP, which is the first of its kind to formulate the automated CPDP model discovery from the perspective of bi-level programming. In particular, the bi-level programming proceeds the optimization with two nested levels in a hierarchical manner. Specifically, the upper-level optimization routine is designed to search for the right combination of transfer learner and classifier while the nested lower-level optimization routine aims to optimize the corresponding hyper-parameter settings. To evaluate BiLO-CPDP, we conduct experiments on 20 projects to compare it with a total of 21 existing CPDP techniques, along with its single-level optimization variant and Auto-Sklearn, a state-of-the-art automated machine learning tool. Empirical results show that BiLO-CPDP champions better prediction performance than all other 21 existing CPDP techniques on 70% of the projects, while being overwhelmingly superior to Auto-Sklearn and its single-level optimization variant on all cases. Furthermore, the unique bi-level formalization in BiLO-CPDP also permits to allocate more budget to the upper-level, which significantly boosts the performance.
@inproceedings{LiXCT20,
author = {Ke Li and
Zilin Xiang and
Tao Chen and
Kay Chen Tan},
title = {BiLO-CPDP},
booktitle = {Proceedings of the 35th IEEE/ACM International Conference on Automated Software Engineering},
year = {2020},
pages = {573--584},
url = {https://ieeexplore.ieee.org/document/9285660},
doi = {10.1145/3324884.3416617}
}
-
Adaptive Operator Selection Based on Dynamic Thompson Sampling for MOEA/D
Lei Sun, Ke Li
Proc. of the 16th International Conference on Parallel Problem Solving from Nature (PPSN XVI), Springer LNCS, volume 12270, p. 271–284, Septermber, 2020.
10.1007/978-3-030-58115-2_19
PDF | BiB | Cited by 29
@inproceedings{SunL20,
author = {Lei Sun and
Ke Li},
title = {Adaptive Operator Selection Based on Dynamic Thompson Sampling for MOEA/D},
booktitle = {Lecture notes in computer science},
year = {2020},
pages = {271--284},
url = {https://link.springer.com/chapter/10.1007%2F978-3-030-58115-2_19},
doi = {10.1007/978-3-030-58115-2_19}
}
-
Surrogate Assisted Evolutionary Algorithm for Medium Scale Multi-Objective Optimisation Problems
Xiaoran Ruan, Ke Li, Bilel Derbel, Arnaud Liefooghe
Proc. of the 22th Annual Conference on Genetic and Evolutionary Computation (GECCO’20), ACM Press: p. 560–568, July, 2020.
10.1145/3377930.3390191
Abs | PDF | BiB | Cited by 21
Building a surrogate model of an objective function has shown to be effective to assist evolutionary algorithms (EAs) to solve real-world complex optimisation problems which involve either computationally expensive numerical simulations or costly physical experiments. However, their effectiveness mostly focuses on small-scale problems with less than 10 decision variables. The scalability of surrogate assisted EAs (SAEAs) have not been well studied yet. In this paper, we propose a Gaussian process surrogate model assisted EA for medium-scale expensive multi-objective optimisation problems with up to 50 decision variables. There are three distinctive features of our proposed SAEA. First, instead of using all decision variables in surrogate model building, we only use those correlated ones to build the surrogate model for each objective function. Second, rather than directly optimising the surrogate objective functions, the original multi-objective optimisation problem is transformed to a new one based on the surrogate models. Last but not the least, a subset selection method is developed to choose a couple of promising candidate solutions for actual objective function evaluations thus to update the training dataset. The effectiveness of our proposed algorithm is validated on benchmark problems with 10, 20, 50 variables, comparing with three state-of-the-art SAEAs.
@inproceedings{RuanLDL20,
author = {Xiaoran Ruan and
Ke Li and
Bilel Derbel and
Arnaud Liefooghe},
title = {Surrogate assisted evolutionary algorithm for medium scale multi-objective optimisation problems},
booktitle = {Proceedings of the 2020 Genetic and Evolutionary Computation Conference},
year = {2020},
pages = {560--568},
url = {https://dl.acm.org/doi/10.1145/3377930.3390191},
doi = {10.1145/3377930.3390191}
}
-
Routing-Led Placement of VNFs in Arbitrary Networks
Joseph Billingsley, Ke Li, Wang Miao, Geyong Min, Nektarios Georgalas
Proc. of the 2020 World Congress on Computational Intelligence (WCCI'20), IEEE Press: p. 1–8, July, 2020.
10.1109/CEC48606.2020.9185531
Abs | PDF | BiB | Cited by 0
The ever increasing demand for computing resources has led to the creation of hyperscale datacentres with tens of thousands of servers. As demand continues to rise, new technologies must be incorporated to ensure high quality services can be provided without the damaging environmental impact of high energy consumption. Virtualisation technology such as network function virtualisation (NFV) allows for the creation of services by connecting component parts known as virtual network functions (VNFs). By optimising the placement and routing of VNFs this technique can be used to maximally utilise available datacentre resources, to maintain a high quality of service whilst minimising energy consumption. Current research on this problem has focussed on placing VNFs and considered routing as a secondary concern. In this work we argue that the opposite approach, a routing-led approach is preferable. We propose a novel routing-led algorithm and analyse each of the component parts over a range of different topologies on problems with up to 16000 variables and compare its performance against a traditional placement based algorithm. Empirical results show that our routing-led algorithm can produce significantly better solutions to large problem instances on a range of datacentre topologies.
@article{BillingsleyLMMG20,
author = {Joseph Billingsley and
Ke Li and
Wang Miao and
Geyong Min and
Nektarios Georgalas},
title = {Routing-Led Placement of VNFs in Arbitrary Networks},
journal = {2020 IEEE Congress on Evolutionary Computation (CEC)},
year = {2020},
volume = {3},
pages = {1--8},
url = {https://ieeexplore.ieee.org/document/9185531},
doi = {10.1109/CEC48606.2020.9185531}
}
-
Surrogate Assisted Evolutionary Algorithm Based on Transfer Learning for Dynamic Expensive Multi-Objective Optimisation Problems
Xuezhou Fan, Ke Li, Kay Chen Tan
Proc. of the 2020 World Congress on Computational Intelligence (WCCI'20), IEEE Press: p. 1–8, July, 2020.
10.1109/CEC48606.2020.9185522
Abs | PDF | BiB | Cited by 18
Dynamic multi-objective optimisation has attracted increasing attention in the evolutionary multi-objective optimisation community in recent years. Comparing to its static counterpart, which has been studied for more than half a century, the involvement of dynamic and uncertain features, including but not limited to the changing Pareto-optimal set, Pareto-optimal front and problem formulation, pose significant more challenges to evolutionary algorithms. This will become even more complicated when the underlying problem involves computationally expensive objective functions which are not rare in many realworld application scenarios. In this paper, we pave an initial step towards the study of dynamic multi-objective optimisation with computationally expensive objective functions. More specifically, we use a surrogate assisted evolutionary algorithm, MOEA/DEGO in particular, as the baseline in order to carry out evolutionary optimisation with a limited amount of function evaluations. Furthermore, instead of restart the MOEA/D-EGO from scratch after each change, we use transfer learning to map the previously archived training data to the current landscape in order to jump start the surrogate model building process. By doing so, we can expect a better adaptation to the new environment. Proof-of-concept experiments fully demonstrate the effectiveness of our proposed method.
@inproceedings{FanLT20,
author = {Xuezhou Fan and
Ke Li and
Kay Chen Tan},
title = {Surrogate Assisted Evolutionary Algorithm Based on Transfer Learning for Dynamic Expensive Multi-Objective Optimisation Problems},
booktitle = {2020 IEEE Congress on Evolutionary Computation (CEC)},
year = {2020},
pages = {1--8},
url = {https://ieeexplore.ieee.org/document/9185522},
doi = {10.1109/CEC48606.2020.9185522}
}
-
DeepSQLi: Deep Semantic Learning for Testing SQL Injection
Muyang Liu, Ke Li, Tao Chen
Proc. of the ACM SIGSOFT 2020 International Symposium on Software Testing and Analysis (ISSTA'20), ACM Press: p. 286–297, July, 2020.
10.1145/3395363.3397375
Abs | PDF | Code | BiB | Cited by 47
Security is unarguably the most serious concern for Web applications, to which SQL injection (SQLi) attack is one of the most devastating attacks. Automatically testing SQLi vulnerabilities is of ultimate importance, yet is unfortunately far from trivial to implement. This is because the existence of a huge, or potentially infinite, number of variants and semantic possibilities of SQL leading to SQLi attacks on various Web applications. In this paper, we propose a deep natural language processing based tool, dubbed DeepSQLi, to generate test cases for detecting SQLi vulnerabilities. Through adopting deep learning based neural language model and sequence of words prediction, DeepSQLi is equipped with the ability to learn the semantic knowledge embedded in SQLi attacks, allowing it to translate user inputs (or a test case) into a new test case, which is se- mantically related and potentially more sophisticated. Experiments are conducted to compare DeepSQLi with SQLmap, a state-of-the-art SQLi testing automation tool, on six real-world Web applications that are of different scales, characteristics and domains. Empirical results demonstrate the effectiveness and the remarkable superiority of DeepSQLi over SQLmap, such that more SQLi vulnerabilities can be identified by using a less number of test cases, whilst running much faster.
@inproceedings{LiuLC20,
author = {Muyang Liu and
Ke Li and
Tao Chen},
title = {DeepSQLi: deep semantic learning for testing SQL injection},
booktitle = {Proceedings of the 29th ACM SIGSOFT International Symposium on Software Testing and Analysis},
year = {2020},
pages = {286--297},
url = {https://dl.acm.org/doi/10.1145/3395363.3397375},
doi = {10.1145/3395363.3397375}
}
-
Understanding the Automated Parameter Optimization on Transfer Learning for Cross-Project Defect Prediction: An Empirical Study
Ke Li*, Zilin Xiang*, Tao Chen*, Shuo Wang, Kay Chen Tan
Proc. of the 42nd International Conference on Software Engineering (ICSE'20), ACM Press: p. 566–577, June, 2020.
10.1145/3377811.3380360
Abs | PDF | Code | BiB | Cited by 54
Data-driven defect prediction has become increasingly important in software engineering process. Since it is not uncommon that data from a software project is insufficient for training a reliable defect prediction model, transfer learning that borrows data/konwledge from other projects to facilitate the model building at the current project, namely cross-project defect prediction (CPDP), is naturally plausible. Most CPDP techniques involve two major steps, i.e., transfer learning and classification, each of which has at least one parameter to be tuned to achieve their optimal performance. This practice fits well with the purpose of automated parameter optimization. However, there is a lack of thorough understanding about what are the impacts of automated parameter optimization on various CPDP techniques. In this paper, we present the first empirical study that looks into such impacts on 62 CPDP techniques, 13 of which are chosen from the existing CPDP literature while the other 49 ones have not been explored before. We build defect prediction models over 20 real-world software projects that are of different scales and characteristics. Our findings demonstrate that: (1) Automated parameter optimization substantially improves the defect prediction performance of 77% CPDP techniques with a manageable computational cost. Thus more efforts on this aspect are required in future CPDP studies. (2) Transfer learning is of ultimate importance in CPDP. Given a tight computational budget, it is more cost-effective to focus on optimizing the parameter configuration of transfer learning algorithms (3) The research on CPDP is far from mature where it is 'not difficult' to find a better alternative by making a combination of existing transfer learning and classification techniques. This finding provides important insights about the future design of CPDP techniques.
@inproceedings{LiX0WT20,
author = {Ke Li and
Zilin Xiang and
Tao Chen and
Shuo Wang and
Kay Chen Tan},
title = {Understanding the automated parameter optimization on transfer learning for cross-project defect prediction},
booktitle = {Proceedings of the ACM/IEEE 42nd International Conference on Software Engineering},
year = {2020},
pages = {566--577},
url = {https://dl.acm.org/doi/abs/10.1145/3377811.3380360},
doi = {10.1145/3377811.3380360}
}
-
On the Combined Impact of Population Size and Sub-problem Selection in MOEA/D
Geoffrey Pruvost, Bilel Derbel, Arnaud Liefooghe, Ke Li, Qingfu Zhang
Proc. of the 20th European Conference on Evolutionary Computation in Combinatorial Optimisation (EvoCOP'20), Springer LNCS, volume 12102, p. 131–147, April, 2020.
10.1007/978-3-030-43680-3_9
PDF | BiB | Cited by 5
@inproceedings{PruvostDLL020,
author = {Geoffrey Pruvost and
Bilel Derbel and
Arnaud Liefooghe and
Ke Li and
Qingfu Zhang},
title = {On the Combined Impact of Population Size and Sub-problem Selection in MOEA/D},
booktitle = {Lecture notes in computer science},
year = {2020},
pages = {131--147},
url = {https://link.springer.com/chapter/10.1007%2F978-3-030-43680-3_9},
doi = {10.1007/978-3-030-43680-3_9}
}
-
Learning to Decompose: A Paradigm for Decomposition-Based Multi-Objective Optimization
Mengyuan Wu, Ke Li, Sam Kwong, Qingfu Zhang, Jun Zhang
IEEE Trans. Evolutionary Computation (TEVC), 23(3): 376–390, 2019.
10.1109/TEVC.2018.2865931
Abs | PDF | Supp | Code | BiB | Cited by 124
The decomposition-based evolutionary multiobjective optimization (EMO) algorithm has become an increasingly popular choice for a posteriori multiobjective optimization. However, recent studies have shown that their performance strongly depends on the Pareto front (PF) shapes. This can be attributed to the decomposition method, of which the reference points and subproblem formulation settings are not well adaptable to various problem characteristics. In this paper, we develop a learning-to-decompose (LTD) paradigm that adaptively sets the decomposition method by learning the characteristics of the estimated PF. Specifically, it consists of two interdependent parts, i.e., a learning module and an optimization module. Given the current nondominated solutions from the optimization module, the learning module periodically learns an analytical model of the estimated PF. Thereafter, useful information is extracted from the learned model to set the decomposition method for the optimization module: 1) reference points compliant with the PF shape and 2) subproblem formulations whose contours and search directions are appropriate for the current status. Accordingly, the optimization module, which can be any decomposition-based EMO algorithm in principle, decomposes the multiobjective optimization problem into a number of subproblems and optimizes them simultaneously. To validate our proposed LTD paradigm, we integrate it with two decomposition-based EMO algorithms, and compare them with four state-of-the-art algorithms on a series of benchmark problems with various PF shapes.
@article{WuLKZZ19,
author = {Mengyuan Wu and
Ke Li and
Sam Kwong and
Qingfu Zhang and
Jun Zhang},
title = {Learning to Decompose: A Paradigm for Decomposition-Based Multiobjective Optimization},
journal = {IEEE Transactions on Evolutionary Computation},
year = {2019},
volume = {23},
number = {3},
pages = {376--390},
url = {https://ieeexplore.ieee.org/document/8439014},
doi = {10.1109/TEVC.2018.2865931}
}
-
Interactive Decomposition Multi-Objective Optimization via Progressively Learned Value Functions
Ke Li*, Renzhi Chen*, Dragan Savic, Xin Yao
IEEE Trans. Fuzzy Systems (TFS), 27(5): 845–860, 2019.
10.1109/TFUZZ.2018.2880700
Abs | PDF | Supp | Code | BiB | Cited by 51
Decomposition has become an increasingly popular technique for evolutionary multiobjective optimization (EMO). A decomposition-based EMO algorithm is usually designed to approximate a whole Pareto-optimal front (PF). However, in practice, a decision maker (DM) might only be concerned in her/his region of interest (ROI), i.e., a part of the PF. Solutions outside that might be useless or even noisy to the decision-making procedure. Furthermore, there is no guarantee that the preferred solutions will be found when many-objective problems. This paper develops an interactive framework for the decomposition-based EMO algorithm to lead a DM to the preferred solutions of her/his choice. It consists of three modules, i.e., consultation, preference elicitation, and optimization. Specifically, after every several generations, the DM is asked to score a few candidate solutions in a consultation session. Thereafter, an approximated value function, which models the DM's preference information, is progressively learned from the DM's behavior. In the preference elicitation session, the preference information learned in the consultation module is translated into the form that can be used in a decomposition-based EMO algorithm, i.e., a set of reference points that are biased toward the ROI. The optimization module, which can be any decomposition-based EMO algorithm in principle, utilizes the biased reference points to guide its search process. Extensive experiments on benchmark problems with three to ten objectives fully demonstrate the effectiveness of our proposed method for finding the DM's preferred solutions.
@article{LiCSY19,
author = {Ke Li and
Renzhi Chen and
Dragan Savic and
Xin Yao},
title = {Interactive Decomposition Multiobjective Optimization Via Progressively Learned Value Functions},
journal = {IEEE Transactions on Fuzzy Systems},
year = {2019},
volume = {27},
number = {5},
pages = {849--860},
url = {https://ieeexplore.ieee.org/document/8531708},
doi = {10.1109/TFUZZ.2018.2880700}
}
-
Two-Archive Evolutionary Algorithm for Constrained Multi-Objective Optimization
Ke Li*, Renzhi Chen*, Guangtao Fu, Xin Yao
IEEE Trans. Evolutionary Computation (TEVC), 23(2): 303–315, 2019.
10.1109/TEVC.2018.2855411
Abs | PDF | Supp | Code | BiB | Cited by 680
When solving constrained multiobjective optimization problems, an important issue is how to balance convergence, diversity, and feasibility simultaneously. To address this issue, this paper proposes a parameter-free constraint handling technique, a two-archive evolutionary algorithm, for constrained multiobjective optimization. It maintains two collaborative archives simultaneously: one, denoted as the convergence-oriented archive (CA), is the driving force to push the population toward the Pareto front; the other one, denoted as the diversity-oriented archive (DA), mainly tends to maintain the population diversity. In particular, to complement the behavior of the CA and provide as much diversified information as possible, the DA aims at exploring areas under-exploited by the CA including the infeasible regions. To leverage the complementary effects of both archives, we develop a restricted mating selection mechanism that adaptively chooses appropriate mating parents from them according to their evolution status. Comprehensive experiments on a series of benchmark problems and a real-world case study fully demonstrate the competitiveness of our proposed algorithm, in comparison to five state-of-the-art constrained evolutionary multiobjective optimizers.
@article{LiCFY19,
author = {Ke Li and
Renzhi Chen and
Guangtao Fu and
Xin Yao},
title = {Two-Archive Evolutionary Algorithm for Constrained Multiobjective Optimization},
journal = {IEEE Transactions on Evolutionary Computation},
year = {2019},
volume = {23},
number = {2},
pages = {303--315},
url = {https://ieeexplore.ieee.org/document/8413136},
doi = {10.1109/TEVC.2018.2855411}
}
-
A Knee-Point-Based Evolutionary Algorithm Using Weighted Subpopulation for Many-Objective Optimization
Juan Zou, Chunhui Ji, Shengxiang Yang, Yuping Zhang, Jinhua Zheng, Ke Li
Swarm and Evolutionary Computation, 47: 33–43, 2019.
10.1016/j.swevo.2019.02.001
PDF | BiB | Cited by 41
@article{ZouJYZZL19,
author = {Juan Zou and
Chunhui Ji and
Shengxiang Yang and
Yuping Zhang and
Jinhua Zheng and
Ke Li},
title = {A knee-point-based evolutionary algorithm using weighted subpopulation for many-objective optimization},
journal = {Swarm and Evolutionary Computation},
year = {2019},
volume = {47},
pages = {33--43},
url = {https://www.sciencedirect.com/science/article/pii/S2210650218300191},
doi = {10.1016/j.swevo.2019.02.001}
}
-
Which Surrogate Works for Empirical Performance Modelling? A Case Study with Differential Evolution
Ke Li, Zilin Xiang, Kay Chen Tan
Proc. of the 2019 IEEE Congress on Evolutionary Computation (CEC'19), IEEE Press: p. 1988–1995, June, 2019.
10.1109/CEC.2019.8789984
Abs | PDF | Supp | BiB | Cited by 13
This is the author accepted manuscript. The final version is available from IEEE via the DOI in this record.
@article{LiXT19,
author = {Ke Li and
Zilin Xiang and
Kay Chen Tan},
title = {Which Surrogate Works for Empirical Performance Modelling? A Case Study with Differential Evolution},
journal = {2019 IEEE Congress on Evolutionary Computation (CEC)},
year = {2019},
pages = {1988--1995},
url = {https://ieeexplore.ieee.org/document/8789984},
doi = {10.1109/CEC.2019.8789984}
}
-
Visualisation of Pareto Front Approximation: A Short Survey and Empirical Comparisons
Huiru Gao, Haifeng Nie, Ke Li
Proc. of the 2019 IEEE Congress on Evolutionary Computation (CEC'19), IEEE Press: p. 1750–1757, June, 2019.
10.1109/CEC.2019.8790298
Abs | PDF | Supp | BiB | Cited by 22
Visualisation is an effective way to facilitate the analysis and understanding of multivariate data. In the context of multi-objective optimisation, comparing to quantitative performance metrics, visualisation is, in principle, able to provide a decision maker better insights about Pareto front approximation sets (e.g. the distribution of solutions, the geometric characteristics of Pareto front approximation) thus to facilitate the decision-making (e.g. the exploration of trade-off relationship, the knee region or region of interest). In this paper, we overview some currently prevalent visualisation techniques according to the way how data is represented. To have a better understanding of the pros and cons of different visualisation techniques, we empirically compare six representative visualisation techniques for the exploratory analysis of different Pareto front approximation sets obtained by four state-of-the-art evolutionary multi-objective optimisation algorithms on the classic DTLZ benchmark test problems. From the empirical results, we find that visual comparisons also follow the No-Free-Lunch theorem where no single visualisation technique is able to provide a comprehensive understanding of the characteristics of a Pareto front approximation set. In other words, a specific type of visualisation technique is only good at exploring a particular aspect of the data.
@article{GaoNL19,
author = {Huiru Gao and
Haifeng Nie and
Ke Li},
title = {Visualisation of Pareto Front Approximation: A Short Survey and Empirical Comparisons},
journal = {2019 IEEE Congress on Evolutionary Computation (CEC)},
year = {2019},
pages = {1750--1757},
url = {https://ieeexplore.ieee.org/abstract/document/8790298},
doi = {10.1109/CEC.2019.8790298}
}
-
Decomposition Multi-Objective Optimisation: Current Developments and Future Opportunities
Ke Li, Qingfu Zhang
Proc. of the 21th Annual Conference on Genetic and Evolutionary Computation (GECCO’19): ACM Press: p. 1002–1031, July 2019.
10.1145/3319619.3323369
Abs | Slides | BiB
tutorial Share on Decomposition multi-objective optimisation: current developments and future opportunities Authors: Ke Li University of Exeter University of ExeterView Profile , Qingfu Zhang City University of Hong Kong City University of Hong KongView Profile Authors Info & Claims GECCO '19: Proceedings of the Genetic and Evolutionary Computation Conference CompanionJuly 2019 Pages 1002–1031https://doi.org/10.1145/3319619.3323369Published:13 July 2019Publication History 1citation148DownloadsMetricsTotal Citations1Total Downloads148Last 12 Months18Last 6 weeks1 Get Citation AlertsNew Citation Alert added!This alert has been successfully added and will be sent to:You will be notified whenever a record that you have chosen has been cited.To manage your alert preferences, click on the button below.Manage my AlertsNew Citation Alert!Please log in to your account Save to BinderSave to BinderCreate a New BinderNameCancelCreateExport CitationPublisher SiteGet Access
@inproceedings{LiZ19,
author = {Ke Li and
Qingfu Zhang},
title = {Decomposition multi-objective optimisation},
booktitle = {Proceedings of the Genetic and Evolutionary Computation Conference Companion},
year = {2019},
pages = {1002--1031},
url = {https://dl.acm.org/citation.cfm?doid=3319619.3323369},
doi = {10.1145/3319619.3323369}
}
-
Security Testing of Web Applications: A Search-Based Approach for Detecting SQL Injection Vulnerabilities
Muyang Liu, Ke Li, Tao Chen
Proc. of the 21th Annual Conference on Genetic and Evolutionary Computation (GECCO’19), ACM Press: p. 417–418, July 2019.
10.1145/3319619.3322026
Abs | PDF | BiB | Cited by 25
Web applications have become increasingly essential in many domains that operate on confidential data related to business. SQL injection attack is one of the most significant web application security risks. Detecting SQL injection vulnerabilities is essential for protecting the underlying web application. However, manually enumerating test cases is extremely challenging, if not impossible, given the potentially infinite number of user inputs and the likely nonexistence of one-to-one mapping between user inputs and malicious SQL statements. This paper proposes an automatic security test case generation approach to detect SQL injection vulnerabilities for web applications, following a search-based software engineering (SBSE) paradigm. Particularly, we propose a novel fitness function that evaluates the similarity between the SQL statements produced by feeding user inputs in the system under test and a known malicious SQL statement. For the search algorithm, we exploit differential evolution, which is robust in continuous optimization but it is under-investigated in SBSE. Based on three real-world web applications, we conduct experiments on 19 configurations that are of diverse forms of SQL statements and types of attacks. Results demonstrate that our approach is more effective, with statistical significance and high effect sizes, than the state-of-the-art.
@inproceedings{LiuLC19,
author = {Muyang Liu and
Ke Li and
Tao Chen},
title = {Security testing of web applications},
booktitle = {Proceedings of the Genetic and Evolutionary Computation Conference Companion},
year = {2019},
pages = {417--418},
url = {https://dl.acm.org/citation.cfm?doid=3319619.3322026},
doi = {10.1145/3319619.3322026}
}
-
Progressive Preference Learning: Proof-of-Principle Results in MOEA/D
Ke Li
Proc. of the 10th International Conference on Evolutionary Multi-Criterion Optimization (EMO’19), Springer LNCS, volume 11411, p. 631–643, March 2019.
10.1007/978-3-030-12598-1_50
PDF | Supp | Code | BiB | Cited by 21
@inproceedings{Li19,
author = {Ke Li},
title = {Progressive Preference Learning: Proof-of-Principle Results in MOEA/D},
booktitle = {Lecture notes in computer science},
year = {2019},
pages = {631--643},
url = {https://link.springer.com/chapter/10.1007%2F978-3-030-12598-1_50},
doi = {10.1007/978-3-030-12598-1_50}
}
-
A Formal Model for Multi-objective Optimisation of NFV Placement
Joseph Billingsley, Ke Li, Wang Miao, Geyong Min, N. Georgalas
Proc. of the 10th International Conference on Evolutionary Multi-Criterion Optimization (EMO’19), Springer LNCS, volume 11411, p. 529–540, March 2019.
10.1007/978-3-030-12598-1_42
PDF | BiB | Cited by 15
@inproceedings{BillingsleyLMMG19,
author = {Joseph Billingsley and
Ke Li and
Wang Miao and
Geyong Min and
Nektarios Georgalas},
title = {A Formal Model for Multi-objective Optimisation of Network Function Virtualisation Placement},
booktitle = {Lecture notes in computer science},
year = {2019},
pages = {529--540},
url = {https://link.springer.com/chapter/10.1007%2F978-3-030-12598-1_42},
doi = {10.1007/978-3-030-12598-1_42}
}
-
R-Metric: Evaluating the Performance of Preference-Based Evolutionary Multi-Objective Optimization Using Reference Points
Ke Li, Kalyanmoy Deb, Xin Yao
IEEE Trans. Evolutionary Computation (TEVC), 22(6): 821–835, 2018.
10.1109/TEVC.2017.2737781
Abs | PDF | Supp | Code | BiB | Cited by 118
Measuring the performance of an algorithm for solving multiobjective optimization problem has always been challenging simply due to two conflicting goals, i.e., convergence and diversity of obtained tradeoff solutions. There are a number of metrics for evaluating the performance of a multiobjective optimizer that approximates the whole Pareto-optimal front. However, for evaluating the quality of a preferred subset of the whole front, the existing metrics are inadequate. In this paper, we suggest a systematic way to adapt the existing metrics to quantitatively evaluate the performance of a preference-based evolutionary multiobjective optimization algorithm using reference points. The basic idea is to preprocess the preferred solution set according to a multicriterion decision making approach before using a regular metric for performance assessment. Extensive experiments on several artificial scenarios, and benchmark problems fully demonstrate its effectiveness in evaluating the quality of different preferred solution sets with regard to various reference points supplied by a decision maker.
@article{LiDY18,
author = {Ke Li and
Kalyanmoy Deb and
Xin Yao},
title = {R-Metric: Evaluating the Performance of Preference-Based Evolutionary Multiobjective Optimization Using Reference Points},
journal = {IEEE Transactions on Evolutionary Computation},
year = {2018},
volume = {22},
number = {6},
pages = {821--835},
url = {https://ieeexplore.ieee.org/document/8049301},
doi = {10.1109/TEVC.2017.2737781}
}
-
Integration of Preferences in Decomposition Multiobjective Optimization
Ke Li*, Renzhi Chen*, Geyong Min, Xin Yao
IEEE Trans. Cybernetics (TCYB), 48(12): 3359–3370, 2018.
10.1109/TCYB.2018.2859363
Abs | PDF | Supp | Code | BiB | Cited by 74
Rather than a whole Pareto-optimal front, which demands too many points (especially in a high-dimensional space), the decision maker (DM) may only be interested in a partial region, called the region of interest (ROI). In this case, solutions outside this region can be noisy to the decision-making procedure. Even worse, there is no guarantee that we can find the preferred solutions when tackling problems with complicated properties or many objectives. In this paper, we develop a systematic way to incorporate the DM's preference information into the decomposition-based evolutionary multiobjective optimization methods. Generally speaking, our basic idea is a nonuniform mapping scheme by which the originally evenly distributed reference points on a canonical simplex can be mapped to new positions close to the aspiration-level vector supplied by the DM. By this means, we are able to steer the search process toward the ROI either directly or interactively and also handle many objectives. Meanwhile, solutions lying on the boundary can be approximated as well given the DM's requirements. Furthermore, the extent of the ROI is intuitively understandable and controllable in a closed form. Extensive experiments on a variety of benchmark problems with 2 to 10 objectives, fully demonstrate the effectiveness of our proposed method for approximating the preferred solutions in the ROI.
@article{LiCMY18,
author = {Ke Li and
Renzhi Chen and
Geyong Min and
Xin Yao},
title = {Integration of Preferences in Decomposition Multiobjective Optimization},
journal = {IEEE Transactions on Cybernetics},
year = {2018},
volume = {48},
number = {12},
pages = {3359--3370},
url = {https://ieeexplore.ieee.org/document/8440670},
doi = {10.1109/TCYB.2018.2859363}
}
-
Evolutionary Multiobjective Optimization-Based Multimodal Optimization: Fitness Landscape Approximation and Peak Detection
Ran Cheng, Miqing Li, Ke Li, Xin Yao
IEEE Trans. Evolutionary Computation (TEVC), 22(5): 692–706, 2018.
10.1109/TEVC.2017.2744328
Abs | PDF | BiB | Cited by 138
Recently, by taking advantage of evolutionary multiobjective optimization techniques in diversity preservation, the means of multiobjectivization has attracted increasing interest in the studies of multimodal optimization (MMO). While most existing work of multiobjectivization aims to find all optimal solutions simultaneously, in this paper, we propose to approximate multimodal fitness landscapes via multiobjectivization, thus providing an estimation of potential optimal areas. To begin with, an MMO problem is transformed into a multiobjective optimization problem (MOP) by adding an adaptive diversity indicator as the second optimization objective, and an approximate fitness landscape is obtained via optimization of the transformed MOP using a multiobjective evolutionary algorithm. Then, on the basis of the approximate fitness landscape, an adaptive peak detection method is proposed to find peaks where optimal solutions may exist. Finally, local search is performed inside the detected peaks on the approximate fitness landscape. To assess the performance of the proposed algorithm, extensive experiments are conducted on 20 multimodal test functions, in comparison with three state-of-the-art algorithms for MMO. Experimental results demonstrate that the proposed algorithm not only shows promising performance in benchmark comparisons, but also has good potential in assisting preference-based decision-making in MMO.
@article{ChengLLY18,
author = {Ran Cheng and
Miqing Li and
Ke Li and
Xin Yao},
title = {Evolutionary Multiobjective Optimization-Based Multimodal Optimization: Fitness Landscape Approximation and Peak Detection},
journal = {IEEE Transactions on Evolutionary Computation},
year = {2018},
volume = {22},
number = {5},
pages = {692--706},
url = {https://ieeexplore.ieee.org/document/8038800},
doi = {10.1109/TEVC.2017.2744328}
}
-
FEMOSAA: Feature Guided and Knee Driven Multi-Objective Optimization for Self-Adaptive Software at Runtime
Tao Chen, Ke Li, Rami Bahsoon, Xin Yao
ACM Trans. Software Engineering and Methodology (TOSEM), 27(2): 1–50, 2018.
10.1145/3204459
Abs | PDF | Code | BiB | Cited by 84
Self-Adaptive Software (SAS) can reconfigure itself to adapt to the changing environment at runtime, aiming to continually optimize conflicted nonfunctional objectives (e.g., response time, energy consumption, throughput, cost, etc.). In this article, we present Feature-guided and knEe-driven Multi-Objective optimization for Self-Adaptive softwAre (FEMOSAA), a novel framework that automatically synergizes the feature model and Multi-Objective Evolutionary Algorithm (MOEA) to optimize SAS at runtime. FEMOSAA operates in two phases: at design time, FEMOSAA automatically transposes the engineers’ design of SAS, expressed as a feature model, to fit the MOEA, creating new chromosome representation and reproduction operators. At runtime, FEMOSAA utilizes the feature model as domain knowledge to guide the search and further extend the MOEA, providing a larger chance for finding better solutions. In addition, we have designed a new method to search for the knee solutions, which can achieve a balanced tradeoff. We comprehensively evaluated FEMOSAA on two running SAS: One is a highly complex SAS with various adaptable real-world software under the realistic workload trace; another is a service-oriented SAS that can be dynamically composed from services. In particular, we compared the effectiveness and overhead of FEMOSAA against four of its variants and three other search-based frameworks for SAS under various scenarios, including three commonly applied MOEAs, two workload patterns, and diverse conflicting quality objectives. The results reveal the effectiveness of FEMOSAA and its superiority over the others with high statistical significance and nontrivial effect sizes.
@article{ChenLBY18,
author = {Tao Chen and
Ke Li and
Rami Bahsoon and
Xin Yao},
title = {FEMOSAA},
journal = {ACM Transactions on Software Engineering and Methodology},
year = {2018},
volume = {27},
number = {2},
pages = {1--50},
url = {https://dl.acm.org/citation.cfm?doid=3234930.3204459},
doi = {10.1145/3204459}
}
-
Dynamic Multi-Objectives Optimization with a Changing Number of Objectives
Ke Li*, Renzhi Chen*, Xin Yao,
IEEE Trans. Evolutionary Computation (TEVC), 21(1): 157–171, 2018.
10.1109/TEVC.2017.2669638
Abs | PDF | Supp | Code | BiB | Cited by 188
Existing studies on dynamic multiobjective optimization (DMO) focus on problems with time-dependent objective functions, while the ones with a changing number of objectives have rarely been considered in the literature. Instead of changing the shape or position of the Pareto-optimal front/set (PF/PS) when having time-dependent objective functions, increasing or decreasing the number of objectives usually leads to the expansion or contraction of the dimension of the PF/PS manifold. Unfortunately, most existing dynamic handling techniques can hardly be adapted to this type of dynamics. In this paper, we report our attempt toward tackling the DMO problems with a changing number of objectives. We implement a dynamic two-archive evolutionary algorithm which maintains two co-evolving populations simultaneously. In particular, these two populations are complementary to each other: one concerns more about the convergence while the other concerns more about the diversity. The compositions of these two populations are adaptively reconstructed once the environment changes. In addition, these two populations interact with each other via a mating selection mechanism. Comprehensive experiments are conducted on various benchmark problems with a time-dependent number of objectives. Empirical results fully demonstrate the effectiveness of our proposed algorithm.
@article{ChenLY18,
author = {Renzhi Chen and
Ke Li and
Xin Yao},
title = {Dynamic Multiobjectives Optimization With a Changing Number of Objectives},
journal = {IEEE Transactions on Evolutionary Computation},
year = {2018},
volume = {22},
number = {1},
pages = {157--171},
url = {https://ieeexplore.ieee.org/document/7886303},
doi = {10.1109/TEVC.2017.2669638}
}
-
Efficient Non-domination Level Update Method for Steady-State Evolutionary Multiobjective Optimization
Ke Li, Kalyanmoy Deb, Qingfu Zhang, Qiang Zhang
IEEE Trans. Cybernetics (TCYB), 47(9): 2838–2849, 2017.
10.1109/TCYB.2016.2621008
Abs | PDF | Supp | Code | BiB | Cited by 101
Nondominated sorting (NDS), which divides a population into several nondomination levels (NDLs), is a basic step in many evolutionary multiobjective optimization (EMO) algorithms. It has been widely studied in a generational evolution model, where the environmental selection is performed after generating a whole population of offspring. However, in a steady-state evolution model, where a population is updated right after the generation of a new candidate, the NDS can be extremely time consuming. This is especially severe when the number of objectives and population size become large. In this paper, we propose an efficient NDL update method to reduce the cost for maintaining the NDL structure in steady-state EMO. Instead of performing the NDS from scratch, our method only updates the NDLs of a limited number of solutions by extracting the knowledge from the current NDL structure. Notice that our NDL update method is performed twice at each iteration. One is after the reproduction, the other is after the environmental selection. Extensive experiments fully demonstrate that, comparing to the other five state-of-the-art NDS methods, our proposed method avoids a significant amount of unnecessary comparisons, not only in the synthetic data sets, but also in some real optimization scenarios. Last but not least, we find that our proposed method is also useful for the generational evolution model.
@article{LiDZZ17,
author = {Ke Li and
Kalyanmoy Deb and
Qingfu Zhang and
Qiang Zhang},
title = {Efficient Nondomination Level Update Method for Steady-State Evolutionary Multiobjective Optimization},
journal = {IEEE Transactions on Cybernetics},
year = {2017},
volume = {47},
number = {9},
pages = {2838--2849},
url = {https://ieeexplore.ieee.org/document/7738460},
doi = {10.1109/TCYB.2016.2621008}
}
-
Matching-Based Selection with Incomplete Lists for Decomposition Multi-Objective Optimization
Mengyuan Wu, Ke Li, Sam Kwong, Yu Zhou, Qingfu Zhang
IEEE Trans. Evolutionary Computation (TEVC), 21(4): 554–568, 2017.
10.1109/TEVC.2017.2656922
Abs | PDF | Supp | Code | BiB | Cited by 82
The balance between convergence and diversity is the cornerstone of evolutionary multiobjective optimization (EMO). The recently proposed stable matching-based selection provides a new perspective to handle this balance under the framework of decomposition multiobjective optimization. In particular, the one-one stable matching between subproblems and solutions, which achieves an equilibrium between their mutual preferences, is claimed to strike a balance between convergence and diversity. However, the original stable marriage model has a high risk of matching a solution with an unfavorable subproblem, which finally leads to an imbalanced selection result. In this paper, we introduce the concept of incomplete preference lists into the stable matching model to remedy the loss of population diversity. In particular, each solution is only allowed to maintain a partial preference list consisting of its favorite subproblems. We implement two versions of stable matching-based selection mechanisms with incomplete preference lists: one achieves a two-level one-one matching and the other obtains a many-one matching. Furthermore, an adaptive mechanism is developed to automatically set the length of the incomplete preference list for each solution according to its local competitiveness. The effectiveness and competitiveness of our proposed methods are validated and compared with several state-of-the-art EMO algorithms on 62 benchmark problems.
@article{WuLKZZ17,
author = {Mengyuan Wu and
Ke Li and
Sam Kwong and
Yu Zhou and
Qingfu Zhang},
title = {Matching-Based Selection With Incomplete Lists for Decomposition Multiobjective Optimization},
journal = {IEEE Transactions on Evolutionary Computation},
year = {2017},
volume = {21},
number = {4},
pages = {554--568},
url = {https://ieeexplore.ieee.org/document/7837621},
doi = {10.1109/TEVC.2017.2656922}
}
-
Recent advances in semantic computing and personalization
Haoran Xie, Fu Lee Wang, Xudong Mao, Ke Li, Qing Li, Handing Wang
Neurocomputing (NEUCOM). 254: 1–2, 2017.
10.1016/j.neucom.2017.02.073
PDF | BiB | Cited by 3
@article{XieWMLLW17,
author = {Haoran Xie and
Fu Lee Wang and
Xudong Mao and
Ke Li and
Qing Li and
Handing Wang},
title = {Recent advances in semantic computing and personalization},
journal = {Neurocomputing},
year = {2017},
volume = {254},
pages = {1--2},
url = {https://www.sciencedirect.com/science/article/pii/S0925231217304058?via%3Dihub},
doi = {10.1016/j.neucom.2017.02.073}
}
-
An Evolutionary Many-Objective Optimization Algorithm Based on Dominance and Decomposition
Ke Li, Kalyanmoy Deb, Qingfu Zhang, Sam Kwong
IEEE Trans. Evolutionary Computation (TEVC), 19(5): 694–716, 2015.
10.1109/TEVC.2014.2373386
Abs | PDF | Code | BiB | Cited by 1197
Achieving balance between convergence and diversity is a key issue in evolutionary multiobjective optimization. Most existing methodologies, which have demonstrated their niche on various practical problems involving two and three objectives, face significant challenges in many-objective optimization. This paper suggests a unified paradigm, which combines dominance- and decomposition-based approaches, for many-objective optimization. Our major purpose is to exploit the merits of both dominance- and decomposition-based approaches to balance the convergence and diversity of the evolutionary process. The performance of our proposed method is validated and compared with four state-of-the-art algorithms on a number of unconstrained benchmark problems with up to 15 objectives. Empirical results fully demonstrate the superiority of our proposed method on all considered test instances. In addition, we extend this method to solve constrained problems having a large number of objectives. Compared to two other recently proposed constrained optimizers, our proposed method shows highly competitive performance on all the constrained optimization problems.
@article{LiDZK15,
author = {Ke Li and
Kalyanmoy Deb and
Qingfu Zhang and
Sam Kwong},
title = {An Evolutionary Many-Objective Optimization Algorithm Based on Dominance and Decomposition},
journal = {IEEE Transactions on Evolutionary Computation},
year = {2015},
volume = {19},
number = {5},
pages = {694--716},
url = {https://ieeexplore.ieee.org/document/6964796},
doi = {10.1109/TEVC.2014.2373386}
}
-
Interrelationship-Based Selection for Decomposition Multiobjective Optimization
Ke Li, Sam Kwong, Qingfu Zhang, Kalyanmoy Deb
IEEE Trans. Cybernetics (TCYB), 45(10): 2076–2088, 2015
10.1109/TCYB.2014.2365354
Abs | PDF | Code | BiB | Cited by 163
Multiobjective evolutionary algorithm based on decomposition (MOEA/D), which bridges the traditional optimization techniques and population-based methods, has become an increasingly popular framework for evolutionary multiobjective optimization. It decomposes a multiobjective optimization problem (MOP) into a number of optimization subproblems. Each subproblem is handled by an agent in a collaborative manner. The selection of MOEA/D is a process of choosing solutions by agents. In particular, each agent has two requirements on its selected solution: one is the convergence toward the efficient front, the other is the distinction with the other agents' choices. This paper suggests addressing these two requirements by defining mutual-preferences between subproblems and solutions. Afterwards, a simple yet effective method is proposed to build an interrelationship between subproblems and solutions, based on their mutual-preferences. At each generation, this interrelationship is used as a guideline to select the elite solutions to survive as the next parents. By considering the mutual-preferences between subproblems and solutions (i.e., the two requirements of each agent), the selection operator is able to balance the convergence and diversity of the search process. Comprehensive experiments are conducted on several MOP test instances with complicated Pareto sets. Empirical results demonstrate the effectiveness and competitiveness of our proposed algorithm.
@article{LiKZD15,
author = {Ke Li and
Sam Kwong and
Qingfu Zhang and
Kalyanmoy Deb},
title = {Interrelationship-Based Selection for Decomposition Multiobjective Optimization},
journal = {IEEE Transactions on Cybernetics},
year = {2015},
volume = {45},
number = {10},
pages = {2076--2088},
url = {https://ieeexplore.ieee.org/document/6975090},
doi = {10.1109/TCYB.2014.2365354}
}
-
A Dual-Population Paradigm for Evolutionary Multiobjective Optimization
Ke Li, Sam Kwong, Kalyanmoy Deb
Information Sciences (INS), 309: 50–72, 2015.
10.1016/j.ins.2015.03.002
PDF | Code | BiB | Cited by 83
@article{LiKD15,
author = {Ke Li and
Sam Kwong and
Kalyanmoy Deb},
title = {A dual-population paradigm for evolutionary multiobjective optimization},
journal = {Information Sciences},
year = {2015},
volume = {309},
pages = {50--72},
url = {https://www.sciencedirect.com/science/article/pii/S0020025515001498?via%3Dihub},
doi = {10.1016/j.ins.2015.03.002}
}
-
Class-Specific Soft Voting based Multiple Extreme Learning Machines Ensemble
Jingjing Cao, Sam Kwong, Ran Wang, Xiaodong Li, Ke Li, Xiangfei Kong
Neurocomputing (NEUCOM). 149: 275–284, 2015.
10.1016/j.neucom.2014.02.072
PDF | BiB | Cited by 76
@article{CaoKWLLK15,
author = {Jingjing Cao and
Sam Kwong and
Ran Wang and
Xiaodong Li and
Ke Li and
Xiangfei Kong},
title = {Class-specific soft voting based multiple extreme learning machines ensemble},
journal = {Neurocomputing},
year = {2015},
volume = {149},
pages = {275--284},
url = {https://www.sciencedirect.com/science/article/pii/S0925231214011345?via%3Dihub},
doi = {10.1016/j.neucom.2014.02.072}
}
-
Stable Matching Based Selection in Evolutionary Multiobjective Optimization
Ke Li, Qingfu Zhang, Sam Kwong, Miqing Li, Ran Wang
IEEE Trans. Evolutionary Computation (TEVC). 18(6): 909–923, 2014.
10.1109/TEVC.2013.2293776
Abs | PDF | Code | BiB | Cited by 360
Multiobjective evolutionary algorithm based on decomposition (MOEA/D) decomposes a multiobjective optimization problem into a set of scalar optimization subproblems and optimizes them in a collaborative manner. Subproblems and solutions are two sets of agents that naturally exist in MOEA/D. The selection of promising solutions for subproblems can be regarded as a matching between subproblems and solutions. Stable matching, proposed in economics, can effectively resolve conflicts of interests among selfish agents in the market. In this paper, we advocate the use of a simple and effective stable matching (STM) model to coordinate the selection process in MOEA/D. In this model, subproblem agents can express their preferences over the solution agents, and vice versa. The stable outcome produced by the STM model matches each subproblem with one single solution, and it tradeoffs convergence and diversity of the evolutionary search. Comprehensive experiments have shown the effectiveness and competitiveness of our MOEA/D algorithm with the STM model. We have also demonstrated that user-preference information can be readily used in our proposed algorithm to find a region that decision makers are interested in.
@article{LiZKLW14,
author = {Ke Li and
Qingfu Zhang and
Sam Kwong and
Miqing Li and
Ran Wang},
title = {Stable Matching-Based Selection in Evolutionary Multiobjective Optimization},
journal = {IEEE Transactions on Evolutionary Computation},
year = {2014},
volume = {18},
number = {6},
pages = {909--923},
url = {https://ieeexplore.ieee.org/document/6678563},
doi = {10.1109/TEVC.2013.2293776}
}
-
Adaptive Operator Selection with Bandits for Multiobjective Evolutionary Algorithm Based on Decomposition
Ke Li, Álvaro Fialho, Sam Kwong, Qingfu Zhang
IEEE Trans. Evolutionary Computation (TEVC). 18(1): 114–130, 2014.
10.1109/TEVC.2013.2239648
Abs | PDF | Supp | Code | BiB | Cited by 406
Adaptive operator selection (AOS) is used to determine the application rates of different operators in an online manner based on their recent performances within an optimization process. This paper proposes a bandit-based AOS method, fitness-rate-rank-based multiarmed bandit (FRRMAB). In order to track the dynamics of the search process, it uses a sliding window to record the recent fitness improvement rates achieved by the operators, while employing a decaying mechanism to increase the selection probability of the best operator. Not much work has been done on AOS in multiobjective evolutionary computation since it is very difficult to measure the fitness improvements quantitatively in most Pareto-dominance-based multiobjective evolutionary algorithms. Multiobjective evolutionary algorithm based on decomposition (MOEA/D) decomposes a multiobjective optimization problem into a number of scalar optimization subproblems and optimizes them simultaneously. Thus, it is natural and feasible to use AOS in MOEA/D. We investigate several important issues in using FRRMAB in MOEA/D. Our experimental results demonstrate that FRRMAB is robust and its operator selection is reasonable. Comparison experiments also indicate that FRRMAB can significantly improve the performance of MOEA/D.
@article{LiFKZ14,
author = {Ke Li and
Alvaro Fialho and
Sam Kwong and
Qingfu Zhang},
title = {Adaptive Operator Selection With Bandits for a Multiobjective Evolutionary Algorithm Based on Decomposition},
journal = {IEEE Transactions on Evolutionary Computation},
year = {2014},
volume = {18},
number = {1},
pages = {114--130},
url = {https://ieeexplore.ieee.org/document/6410018},
doi = {10.1109/TEVC.2013.2239648}
}
-
A General Framework for Evolutionary Multiobjective Optimization via Manifold Learning
Ke Li, Sam Kwong
Neurocomputing (NEUCOM). 146: 65–74, 2014.
10.1016/j.neucom.2014.03.070
PDF | BiB | Cited by 53
@article{LiK14,
author = {Ke Li and
Sam Kwong},
title = {A general framework for evolutionary multiobjective optimization via manifold learning},
journal = {Neurocomputing},
year = {2014},
volume = {146},
pages = {65--74},
url = {https://www.sciencedirect.com/science/article/pii/S0925231214008686?via%3Dihub},
doi = {10.1016/j.neucom.2014.03.070}
}
-
Evolutionary Algorithms with Segment-based Search for Multiobjective Optimization Problems
Miqing Li, Shengxiang Yang, Ke Li, Xiaohui Liu
IEEE Trans. Cybernetics (TCYB). 44(8): 1295–1313, 2014.
10.1109/TCYB.2013.2282503
Abs | PDF | BiB | Cited by 44
This paper proposes a variation operator, called segment-based search (SBS), to improve the performance of evolutionary algorithms on continuous multiobjective optimization problems. SBS divides the search space into many small segments according to the evolutionary information feedback from the set of current optimal solutions. Two operations, micro-jumping and macro-jumping, are implemented upon these segments in order to guide an efficient information exchange among "good" individuals. Moreover, the running of SBS is adaptive according to the current evolutionary status. SBS is activated only when the population evolves slowly, depending on general genetic operators (e.g., mutation and crossover). A comprehensive set of 36 test problems is employed for experimental verification. The influence of two algorithm settings (i.e., the dimensionality and boundary relaxation strategy) and two probability parameters in SBS (i.e., the SBS rate and micro-jumping proportion) are investigated in detail. Moreover, an empirical comparative study with three representative variation operators is carried out. Experimental results show that the incorporation of SBS into the optimization process can improve the performance of evolutionary algorithms for multiobjective optimization problems.
@article{LiYLL14,
author = {Miqing Li and
Shengxiang Yang and
Ke Li and
Xiaohui Liu},
title = {Evolutionary Algorithms With Segment-Based Search for Multiobjective Optimization Problems},
journal = {IEEE Transactions on Cybernetics},
year = {2014},
volume = {44},
number = {8},
pages = {1295--1313},
url = {https://ieeexplore.ieee.org/document/6627937},
doi = {10.1109/TCYB.2013.2282503}
}
-
A Novel Algorithm for Non-dominated Hypervolume-based Multiobjective Optimization
Ke Li, Jinhua Zheng, Miqing Li, Cong Zhou, Hui Lv
Proc. of 2009 IEEE International Conference on Systems, Mans and Cybernetics (SMC’09), IEEE Press: p. 5220–5226, December 2009.
10.1109/ICSMC.2009.5345983
Abs | PDF | BiB | Cited by 13
Hypervolume indicator is a commonly accepted quality measure to assess the set of non-dominated solutions obtained by an evolutionary multiobjective optimization algorithm. Recently, an emerging trend in the design of evolutionary multiobjective optimization algorithms is to directly optimize a quality indicator. In this paper, we propose a hypervolume-based evolutionary algorithm for multiobjective optimization. There are two main contributions of our approach, on one hand, a unique fitness assignment strategy is proposed, on the other hand, we design a slicing based method to calculate the exclusive hypervolume of each individual for environmental selection. From an extensive comparative study with three other MOEAs on a number of two and three objective test problems, it is observed that the proposed algorithm has good performance in convergence and distribution.
@inproceedings{LiZLZL09,
author = {Ke Li and
Jinhua Zheng and
Miqing Li and
Cong Zhou and
Hui Lv},
title = {A novel algorithm for non-dominated hypervolume-based multiobjective optimization},
booktitle = {2009 IEEE International Conference on Systems, Man and Cybernetics},
year = {2009},
pages = {5220--5226},
url = {https://ieeexplore.ieee.org/document/5345983},
doi = {10.1109/ICSMC.2009.5345983}
}
-
An Spanning Tree Based Method For Pruning Non-Dominated Solutions in Multi- Objective Optimization Problems
Miqing Li, Jinhua Zheng, Ke Li, Jun Wu, Guixia Xiao
Proc. of 2009 IEEE International Conference on Systems, Mans and Cybernetics (SMC’09), IEEE Press: p. 4882–4887, December 2009.
10.1109/ICSMC.2009.5346322
Abs | PDF | BiB | Cited by 2
Diversity maintenance of solutions is a crucial part in multi-objective optimization. However, most of existing studies show a good distribution with a large computational load or a comparative bad distribution quickly. In this paper, a method for pruning a set of non-dominated solutions using a Spanning Tree is proposed. This approach defines a density estimation metric — Spanning Tree Crowding Distance (STCD). Moreover, information of degree of solution combined with STCD is employed to truncate the population. From an extensive comparative study with three other methods on a number of 2, 3 and 4 objective test problems, the proposed method indicates a good balance among uniformity, spread and execution time.
@inproceedings{LiZLWX09,
author = {Miqing Li and
Jinhua Zheng and
Ke Li and
Jun Wu and
Guixia Xiao},
title = {An Spanning Tree based method for pruning non-dominated solutions in multi-objective optimization problems},
booktitle = {2009 IEEE International Conference on Systems, Man and Cybernetics},
year = {2009},
pages = {4882--4887},
url = {https://ieeexplore.ieee.org/document/5346322/},
doi = {10.1109/ICSMC.2009.5346322}
}
-
Objective Reduction based on the Least Square Method for Large-dimensional Multiobjective Optimization Problem
Cong Zhou, Jinhua Zheng, Ke Li, Hui Lv
Proc. of the 5th International Conference on Natural Computation (ICNC’09), IEEE Press: p. 350–354, August 2009.
10.1109/ICNC.2009.40
Abs | PDF | BiB | Cited by 10
In the real-world applications, many multi-objective optimization involve a large number of objective, however, existing evolutionary multi-objective optimization algorithms are applied only to a few number of objective. Because of inconvenience in handling large number of objective, researchers start to deal with how to reduce the redundant objectives. In this paper, we firstly introduce some existing algorithms on transforming high-dimensional to low-dimensional, and then propose a new algorithm, namely large dimensionality reduction based on the least square method. This method fits every objective function to a line, and compares the slope differences between each two lines, finally makes certain which one is redundancy and further reduces this one. This experiment shows, on one hand, there are some redundant objective functions in certain large dimensionality multi-objective optimization problems, and the objective space of non-redundant objective function is accordant with the low-dimensional true Pareto front. On other hand, the experiment result with other similar algorithm shows our algorithm is competitive and the efficacy of the procedure is demonstrated.
@inproceedings{ZhouZLL09,
author = {Cong Zhou and
Jinhua Zheng and
Ke Li and
Hui Lv},
title = {Objective Reduction Based on the Least Square Method for Large-Dimensional Multi-objective Optimization Problem},
booktitle = {2009 Fifth International Conference on Natural Computation},
year = {2009},
pages = {350--354},
url = {https://ieeexplore.ieee.org/document/5366350},
doi = {10.1109/ICNC.2009.40}
}
-
The Convergence Analysis of Genetic Algorithm based on Space Mating
Hui Lv, Jinhua Zheng, Jun Wu, Cong Zhou, Ke Li
Proc. of the 5th International Conference on Natural Computation (ICNC’09), IEEE Press: p. 557–562, August 2009.
10.1109/ICNC.2009.39
Abs | PDF | BiB | Cited by 4
This paper analyzes the convergence properties of the genetic algorithm based on space mating with mutation, crossover and proportional reproduction applied to static optimization on problems. It is proved by means of homogeneous finite Markov chain analysis that genetic algorithm based on space mating will converge to the global optimum. Each process is convergence to the global optimum, at least satisfactory solution under the best individual survives besides the last course. And illuminate a population converge with probability one in the no mutation operator conditions. By comparing the experiment, we can see that the algorithm have better convergence than SGA and consist with the theory.
@inproceedings{LvZWZL09,
author = {Hui Lv and
Jinghua Zheng and
Jun Wu and
Cong Zhou and
Ke Li},
title = {The Convergence Analysis of Genetic Algorithm Based on Space Mating},
booktitle = {2009 Fifth International Conference on Natural Computation},
year = {2009},
volume = {16},
pages = {557--562},
url = {https://ieeexplore.ieee.org/document/5366015},
doi = {10.1109/ICNC.2009.39}
}
-
An Improved Multi-objective Evolutionary Algorithm based on Differential Evolution
Ke Li, Jinhua Zheng
Proc. of 2009 WRI World Congress on Computer Science and Information Engineering (CSIE’09), IEEE Press: p. 825–830, March 2009.
10.1109/CSIE.2009.181
Abs | PDF | BiB | Cited by 13
Evolutionary algorithms (EAs) are well-known optimization approaches to deal with nonlinear and complex problems. However, these population-based algorithms are computationally expensive due to the slow nature of the evolutionary process. This paper proposes an improved differential evolution algorithm (CDE). On the one hand CDE combines the advantages of DE with the mechanisms of Pareto based ranking and crowding distance sorting which are similar to the NSGA-II, on the other hand different from the previous DE, CDE compares the trial vector to its nearest neighbor to decide whether to preserve it. Experimental results confirm that CDE outperforms the other two classical multi-objective evolutionary algorithms (MOEAs) NSGA-II and SPEA2 in terms of diversity and convergence
@inproceedings{LiZZL09,
author = {Ke Li and
Jinhua Zheng and
Cong Zhou and
Hui Lv},
title = {An Improved Differential Evolution for Multi-objective Optimization},
booktitle = {2009 WRI World Congress on Computer Science and Information Engineering},
year = {2009},
pages = {825--830},
url = {https://ieeexplore.ieee.org/document/5171111},
doi = {10.1109/CSIE.2009.181}
}