Shengcai LIU (刘晟材) 助理教授，博导 Tenure-Track Assistant Professor, Ph.D Supervisor Department of Computer Science and Engineering Southern University of Science and Technology Office: Room 311, South Tower, College of Engineering Address: No. 1088 Xueyuan Avenue, Shenzhen, China 518055 Email: liusc3 AT sustech DOT edu DOT cn

Short Biography

I received my B.Sc. and Ph.D. degrees in the School of Computer Science and Technology from USTC in 2014 and 2020, respectively, advised by Prof. Xin Yao and Prof. Ke Tang. From Jan 2021 to Jan 2023, I was a Research Assistant Professor at the CSE Department of SUSTech. From Jan 2023 to May 2024, I was a Visiting Professor, and later a senior scientist, at CFAR, A*STAR, Singapore, in collaboration with Prof. Yew-Soon Ong from NTU. I am also a member of the Nature Inspired Computation and Applications Laboratory (NICAL) led by Prof. Xin Yao and Prof. Ke Tang.

个人简介

刘晟材博士，分别在2014年和2020年于中国科学技术大学计算机学院获得学士和博士学位。曾在新加坡科技研究局担任高级研究科学家，现任南方科技大学计算机科学与工程系助理教授、博导，北京中关村学院兼职导师。长期从事学习优化 (Learn to optimize)、智能优化基座大模型、AI for Material等方向的研究。已获批主持国自然基金一项、华为合作项目两项，迄今已在NeurIPS、ICML、AAAI等CCF-A类会议以及IEEE TEVC、IEEE TCYB等智能优化旗舰期刊上以一作/通讯作者身份发表论文20余篇。刘博士曾多次受邀在国际学术会议上做大会报告，包括达格施图尔城堡研讨会2023、IEEE世界计算智能大会2022、以及第七届演化计算与学习研讨会。

Research Interests

Ultimately, I am obsessed with the fully-automated design of optimizers, as well as their applications in high-impact applications. Now I am interested in:

• Learning easy-to-use, general-purpose optimizers for discrete optimization problems
• Learning optimizers for material design
• Learning optimizers for post training of large language models (LLMs)
• LLMs for Optimization

招募: 本课题组常年招收硕士/博士研究生、访问学生、博士后、研究助理教授，不定时招收北京中关村学院-南方科技大学联合培养博士生。 欢迎数理基础扎实、代码能力强的同学申请。申请时请将本人自述、成绩单、个人简历及相关能力发送至邮箱 liusc3 AT sustech.edu.cn 以及 chengy6 AT mail.sustech.edu.cn，邮件主题请注明为 申请人姓名+申请xx职位/xxx年入学博士生/硕士生.

Recruitment: We have open positions for M.Sc./Ph.D. students, Visiting Scholars, Postdocs, and Research Assistant Professors. Candidates with strong backgrounds in mathematics and programming are highly encouraged to apply. To Apply: Please email your CV, transcripts, and personal statement to liusc3 AT sustech.edu.cn and chengy6 AT mail.sustech.edu.cn. Subject Line: Name + Position Applied / Intake Year.

Grants & Awards

• 多特性弧路径规划问题的求解方法研究, 国家自然科学基金青年科学基金项目（C类）, 2026/01-2028/12, PI
• 大模型RL微调的XXX技术, 华为技术有限公司, 2025/10-2026/10, PI
• XXX求解器的自动构建关键方法, 华为技术有限公司, 2021/05-2022/05, PI

Selected Publications

Preprints (First/corresponding-authored works listed; See Google Scholar for the complete list.)

1. Zhiyuan Wang, Shengcai Liu, Shaofeng Zhang, Ke Tang. A Novel Population Initialization Method via Adaptive Experience Transfer for General-Purpose Binary Evolutionary Optimization.
   arXiv preprint arXiv:2512.00341. [Arxiv]
   
   
2. Wenjie Chen, Li Zhuang, Ziying Luo, Yu Liu, Jiahao Wu, Shengcai Liu. Personalized Treatment Outcome Prediction from Scarce Data via Dual-Channel Knowledge Distillation and Adaptive Fusion.
   arXiv preprint arXiv:2510.26444. [Arxiv]
   
   
3. Shengcai Liu, Hui Ou-yang, Zhiyuan Wang, Cheng Chen, Qijun Cai, Yew-Soon Ong, Ke Tang. Scalable Structure Learning of Bayesian Networks by Learning Algorithm Ensembles.
   arXiv preprint arXiv:2506.22848. [Arxiv]
   
   
4. Shaofeng Zhang, Shengcai Liu, Ning Lu, Jiahao Wu, Ji Liu, Yew-Soon Ong, Ke Tang. LLM-Driven Instance-Specific Heuristic Generation and Selection.
   arXiv preprint arXiv:2506.00490. [Arxiv]
   
   
5. Zeyu Dai, Shengcai Liu, Rui He, Jiahao Wu, Ning Lu, Wenqi Fan, Qing Li, and Ke Tang. SemDiff: Generating Natural Unrestricted Adversarial Examples via Semantic Attributes Optimization in Diffusion Models.
   Arxiv preprint arXiv:2504.11923. [Arxiv]
   
   
6. Shengcai Liu, Zhiyuan Wang, Yew-Soon Ong, Xin Yao, and Ke Tang. Learning Mixture-of-Experts for General-Purpose Black-Box Discrete Optimization.
   Arxiv preprint arXiv:2405.18884. [Arxiv]
   
   

Journal Papers

1. Rui Xu, Xing Fan, Shengcai Liu*, Wenjie Chen, Ke Tang. Memetic Search for Green Vehicle Routing Problem with Private Capacitated Refueling Stations.
   IEEE Transactions on Evolutionary Computation, 2025, To appear. [Paper] [Arxiv] [Code]
   
   
2. Zhiyuan Wang, Shengcai Liu*, Peng Yang, and Ke Tang. Evolving Generalizable Parallel Algorithm Portfolios for Binary Optimization Problems via Domain-Agnostic Instance Generation.
   IEEE Transactions on Evolutionary Computation, 2025, To appear. [Paper] [Arxiv] [Code]
   
   
3. Wenjie Chen, Shengcai Liu*, Yew-Soon Ong, Li Zhuang, and Ke Tang. Neural Influence Estimator: Towards Real-time Solutions to Influence Blocking Maximization.
   IEEE Transactions on Computational Social Systems, 2025, 12(6): 5155-5167. [Paper] [Arxiv]
   
   
4. Zubin Zheng, Shengcai Liu*, and Yew-Soon Ong. Hybrid Memetic Search for Electric Vehicle Routing with Time Windows, Simultaneous Pickup-Delivery, and Partial Recharges.
   IEEE Transactions on Emerging Topics in Computational Intelligence, 2025, 9(6): 3773-3787. [Paper] [Arxiv] [Code]
   
   
5. Jiahao Wu, Wenqi Fan, Jingfan Chen, Shengcai Liu*, Qijiong Liu, Rui He, Qing Li, and Ke Tang. Condensing Pre-augmented Recommendation Data via Lightweight Policy Gradient Estimation.
   IEEE Transactions on Knowledge and Data Engineering, 2025, 37(1): 162 - 173. [Paper] [Arxiv] [Code]
   
   
6. Xuanfeng Li, Shengcai Liu*, and Ke Tang. Novel Genetic Algorithm for Solving Chance-Constrained Multiple-Choice Knapsack Problems.
   Journal of Computer Applications, 2024, 44(5): 1378-1385. [Paper]
   
   
7. Xuanfeng Li, Shengcai Liu*, Jin Wang, Xiao Chen, Yew-Soon Ong, and Ke Tang. Chance-Constrained Multiple-Choice Knapsack Problem: Model, Algorithms, and Applications.
   IEEE Transactions on Cybernetics, 2024, 54(12): 7969 - 7980. [Paper] [Arxiv] [Code]
   
   
8. Ning Lu, Shengcai Liu*, Rui He, Qi Wang, Yew-Soon Ong, and Ke Tang. Large Language Models can be Guided to Evade AI-Generated Text Detection.
   Transactions on Machine Learning Research, 2024. [Paper] [Arxiv] [Code]
   
   
9. Shengcai Liu, Ning Lu, Wenjing Hong, Chao Qian, and Ke Tang. Effective and Imperceptible Adversarial Textual Attack via Multi-objectivization.
   ACM Transactions on Evolutionary Learning and Optimization, 2024, 4(3): 16:1-16:23. [Paper] [Arxiv] [Code]
   
   
10. Shengcai Liu, Yu Zhang, Ke Tang, and Xin Yao. How Good Is Neural Combinatorial Optimization? A Systematic Evaluation on the Traveling Salesman Problem.
    IEEE Computational Intelligence Magazine, 2023, 18(3): 14-28. [Paper] [Arxiv] [Code]
    
    
11. Zeyu Dai, Shengcai Liu*, Qing Li, and Ke Tang. Saliency Attack: Towards Imperceptible Black-box Adversarial Attack.
    ACM Transactions on Intelligent Systems and Technology, 2023, 14(3): 1-20. [Paper] [Arxiv] [Code]
    
    
12. Rui He, Shengcai Liu*, Shan He, and Ke Tang. Multi-Domain Active Learning: Literature Review and Comparative Study.
    IEEE Transactions on Emerging Topics in Computational Intelligence, 2023, 7(3): 791-804. [Paper] [Arxiv]
    
    
13. Shengcai Liu, Ning Lu, Cheng Chen, and Ke Tang. Efficient Combinatorial Optimization for Word-level Adversarial Textual Attack.
    IEEE/ACM Transactions on Audio, Speech and Language Processing, 2022, 30: 98-111. [Paper] [Arxiv] [Code]
    
    
14. Shengcai Liu, Peng Yang, and Ke Tang. Approximately Optimal Construction of Parallel Algorithm Portfolios by Evolutionary Intelligence (in Chinese).
    SCIENTIA SINICA Technologica, 2023, 53(2): 280-290. [Paper]
    
    
15. Shengcai Liu, Ke Tang, and Xin Yao. Generative Adversarial Construction of Parallel Portfolios.
    IEEE Transactions on Cybernetics, 2022, 52(2): 784-795. [Paper]
    
    
16. Shengcai Liu, Ke Tang, and Xin Yao. Memetic Search for Vehicle Routing with Simultaneous Pickup-Delivery and Time Windows.
    Swarm and Evolutionary Computation, 66: 100927, 2021. [Paper] [Arxiv] [Code]
    
    
17. Shengcai Liu, Ke Tang, Peng Yang, and Xin Yao. Few-shots Parallel Algorithm Portfolio Construction via Co-evolution.
    IEEE Transactions on Evolutionary Computation, 2021, 25(3): 595-607. [Paper] [Arxiv] [Code]
    
    

Conference Papers

1. Zhangying Feng, Qianglong Chen, Ning Lu, Yongqian Li, Siqi Cheng, Shuangmu Peng, Duyu Tang, Shengcai Liu*, Zhirui Zhang*. Is PRM Necessary? Problem-Solving RL Implicitly Induces PRM Capability in LLMs.
   In: Proceedings of the 39th Annual Conference on Neural Information Processing Systems (NeurIPS'2025). [Arxiv]
   
   
2. Haoze Lv, Zhiyuan Wang, Wenjie Chen, Shengcai Liu*. Cascaded Large-Scale TSP Solving with Unified Neural Guidance: Bridging Local and Population-based Search.
   In: Proceedings of the 28th European Conference on Artificial Intelligence (ECAI'2025), Bologna, Italy, 2025. [Arxiv]
   
   
3. Ning Lu, Shengcai Liu*, Jiahao Wu, Weiyu Chen, Zhirui Zhang, Yew-Soon Ong, Qi Wang, Ke Tang. Safe Delta: Consistently Preserving Safety when Fine-Tuning LLMs on Diverse Datasets.
   In: Proceedings of the 42nd International Conference on Machine Learning (ICML'2025), Vancouver, Canada, 2025. [Paper] [Arxiv] [Code]
   
   
4. Jiahao Wu, Ning Lu, Zeiyu Dai, Wenqi Fan, Shengcai Liu*, Qing Li, Ke Tang. Backdoor Graph Condensation.
   In: Proceedings of The 40th IEEE International Conference on Data Engineering (ICDE'2025), HongKong, China, 2025, 2267-2280. [Paper] [Arxiv] [Code]
   
   
5. Jiahao Wu, Qujiong Liu, Hengchang Hu, Wenqi Fan, Shengcai Liu*, Qing Li, Xiao-Ming Wu, Ke Tang. Leveraging ChatGPT to Empower Training-free Dataset Condensation for Content based Recommendation.
   In: Companion Proceedings of the ACM on Web Conference 2025 (WWW'2025), Sydney, Australia, 2025, 1402-1406. [Paper] [Arxiv] [Code]
   
   
6. Ning Lu, Shengcai Liu*, Zhirui Zhang, Qi Wang, Haifeng Liu, and Ke Tang. Less is More: Understanding Word-level Textual Adversarial Attack via n-gram Frequency Descend.
   In: Proceedings of The 2024 IEEE Conference on Artificial Intelligence (CAI'2024), Singapore, Singapore, 2024, 823-830. [Paper] [Arxiv]
   
   
7. Shengcai Liu, Caishun Chen, Xinghua Qu, Ke Tang, and Yew-Soon Ong. Large Language Models as Evolutionary Optimizers.
   In: Proceedings of The 2024 IEEE Congress on Evolutionary Computation (CEC'2024), Yokohama, Japan, 2024, 1-8. [Paper] [Arxiv] [Code]
   
   
8. Rui He, Shengcai Liu*, Jiahao Wu, Shan He, and Ke Tang. Multi-Domain Learning From Insufficient Annotations.
   In: Proceedings of The 26th European Conference on Artificial Intelligence (ECAI'2023), Kraków, Poland, 2023, 1028-1035. [Paper] [Arxiv]
   
   
9. Shengcai Liu, Fu Peng, and Ke Tang. Reliable Robustness Evaluation via Automatically Constructed Attack Ensembles.
   In: Proceedings of The 37th AAAI Conference on Artificial Intelligence (AAAI'2023), Washington, DC, 2023, 8852-8860. [Paper] [Arxiv] [Code]
   
   
10. Fu Peng, Shengcai Liu*, and Ke Tang. Training Quantized Deep Neural Networks via Cooperative Coevolution.
    In: Proceedings of the 13th International Conference on Swarm Intelligence (ICSI'2022), Xi'an, China, 2022, 81-93. [Paper]
    
    
11. Kangfei Zhao, Shengcai Liu*, Yu Rong, and Jeffrey Xu Yu. Towards Feature-free TSP Solver Selection: A Deep Learning Approach.
    In: Proceedings of the 20th International Joint Conference on Neural Networks (IJCNN'2021), Virtual Event, 2021, 1-8. [Paper] [Arxiv] [Code]
    
    
12. Shengcai Liu, Ke Tang, and Xin Yao. On Performance Estimation in Automatic Algorithm Configuration.
    In: Proceedings of The 34th AAAI Conference on Artificial Intelligence (AAAI'2020), New York, NY, 2020, 2384-2391. [Paper]
    
    
13. Shengcai Liu, Ke Tang, and Xin Yao. Automatic Construction of Parallel Portfolios via Explicit Instance Grouping.
    In: Proceedings of The 33rd AAAI Conference on Artificial Intelligence (AAAI'2019), Honululu, HI, 2019, 1560-1567. [Paper]
    
    
14. Shengcai Liu, Yufan Wei, Ke Tang, A.K. Qin, and Xin Yao. Qos-aware Long-Term Based Service Composition in Cloud Computing.
    In: Proceedings of The 14th IEEE Congress on Evolutionary Computation (CEC'2015), Sendai, Japan, 2015, 3362-3369. [Paper]
    
    

Professional Activities

• Academic Participation:

北京中关村学院兼职导师 (2025-)
IEEE Member (2022-)


• Invited Talks:

“优化算法的自动生成”, 2025湖南人工智能大会-智能系统与优化论坛, 2025年11月22日, 湘潭.
“Some Preliminary Attempts on LLM-Driven Optimization and Optimization for LLMs”, 第4届智能决策论坛, 2025年6月15日, 南京.
“迈向通用离散优化智能体”, 第20次南澳科学会议, 2025年4月27日, 汕头.
“Large Language Models as General-Purpose Evolutionary Optimizers”, 2024 Workshop on Learning-to-Optimize Paradigms in the Era of LLMs, 2024年12月3日, 深圳华为.
“How Good is Neural Combinatorial Optimization?”, Dagstuhl Seminar “Synergizing Theory and Practice of Automated Algorithm Design for Optimization”. Aug 17, 2023. Dagstuhl, Germany.
“Learn to Optimize”, The 2022 IEEE World Congress on Computational Intelligence (WCCI’2022). Jul 18, 2022. [Slides], Online.
“Co-Evolved Parallel Algorithm Portfolios”, 第七届演化计算与学习研讨会 (The 7th Workshop on Evolutionary Computation and Learning, ECOLE’2021), 2021年5月15日, 武汉.

Teaching

• CSE5025 - Combinatorial Optimization (组合优化), for graduate students, 2025 Fall
• CS109 - Introduction to Java Programming, for undergraduate students, 2024 Fall