Young Scholars Forum
Ye Tian
Associate Professor
Anhui University, China
Bio: Ye Tian is now an associate professor in Anhui University. His interests include evolutionary computation and deep learning. He has published more than 50 papers on more than ten IEEE Trans with more than 10000 citations. He is a Highly Cited Researcher of Clarivate and a Most Cited Chinese Researcher of Elsevier in 2023. He is the recipient of the 2018, 2021, and 2024 IEEE TEVC Outstanding Paper Awards, the 2020 IEEE CIM Outstanding Paper Award, the 2022 IEEE CIS Outstanding PhD Dissertation Award, and the 2024 IEEE TEVC Outstanding Associate Editor. He is the author of PlatEMO.
The Speech title: Automated creation of evolutionary algorithms from scratch
Abstract: In recent years, numerous evolutionary algorithms have been proposed to solve various optimization problems. These evolutionary algorithms are mostly manually designed, based on common modules such as genetic operators and differential evolution operators. In this presentation, we propose an automated evolutionary algorithm creation framework that can generate entirely new evolutionary algorithms from scratch. Experiments have shown that the algorithms created by this framework outperform existing algorithms on single-objective, multi-objective, and combinatorial optimization problems. Moreover, they even outperform surrogate-assisted algorithms on expensive optimization problems without using surrogate models. Besides, we have developed a new GUI in PlatEMO v4.8, allowing users to build their own evolutionary algorithms without writing any code and easily surpass existing evolutionary algorithms.
Cheng He
Associate Professor
Huazhong University of Science and Technology, China
Bio: Dr. Cheng He is currently an Associate Professor with the School of Electrical and Electronic Engineering, Huazhong University of Science and Technology, China. His main research interests are Artificial/Computational Intelligence and its Applications (including evolutionary multi-objective optimization, model-based optimization, large-scale optimization, etc.).
He has published over 40 SCI papers, including 4 ESI highly citation papers, and 19 IEEE TEVC, TCYB, TETCI, TSMCS, TNNLS, and TAI).
He is an IEEE Senior Member, the Associate Editor of Complex & Intelligent Systems, Editorial Board members of PloS One and Electronics, the Chair of the IEEE CIS Taskforce on Intelligence Systems for Health, and the recipient of the 2021 ACM China Council SIGBIO Chapter Rising Star Award. He has organized three Special Issues on different journals, over ten Special Sessions and three competitions on IEEE CEC/SSCI/WCCI.
The Speech title: Evolutionary Large-scale Multiobjective Optimization: Challenges and Opportunities
Abstract: Evolutionary algorithms (EAs) have been a popular optimization tool for decades, which have shown promising performance in solving various benchmark optimization problems. Nevertheless, using EAs on problems with over 100 decision variables (large-scale optimization problems, LSOPs) remains challenging due to the "curse of dimensionality", especially for those LSOPs in real-world applications. In this presentation, some EAs for handling the challenges regarding the computational efficiency, balanced performance, and constraints are first introduced. Then, some real-world applications of modelling and solving LSMOPs are explicated. Eventually, some open questions regarding the opportunities of EAs with the assistance of advanced learning models (such as physics-informed neural network and large language models) are discussed.
Ruwang Jiao
Associate Professor
Soochow University, China
Bio: Ruwang Jiao is an associate professor at Soochow University, China. He was a research fellow in artificial intelligence at Victoria University of Wellington, New Zealand. He received the Humboldt Research Fellowship in 2024 and currently chairs the IEEE CIS Task Force on Evolutionary Computation for Feature Selection and Construction. He co-chaired the special session on EMOML at IEEE CEC 2023 and IEEE WCCI 2024 and delivered a tutorial at IEEE SSCI 2023. His research mainly focuses on feature selection, dimensionality reduction, and evolutionary multi-objective learning. He has published over 30 papers in esteemed journals and conferences, including IEEE TEVC, ECJ (MIT Press), IEEE TCYB, and IEEE TAP.
The Speech title: Evolutionary Multi-objective Feature Selection for Machine Learning
Abstract: Maximizing classification accuracy and minimizing the number of selected features are the two primary objectives in feature selection, making it a multi-objective task. Multi-objective feature selection not only reduces dimensionality and improves accuracy, but also provides insights from complex data. Despite significant advancements, there are still some challenges that need to be addressed, such as complex feature interactions, exponentially growing search space, solution duplication in the objective space, and objective selection bias. In this talk, we will introduce recent works addressing these challenges and highlight emerging directions for the future of multi-objective feature selection.
Qianlong Dang
Associate Professor
Northwest A & F University, China
Bio: Qianlong Dang, Associate Professor/Master’s Supervisor of Mathematics Department at Northwest A&F University. Obtained a doctoral degree from the School of Mathematics and Statistics at Xidian University in 2022. My research interest lies in the study of artificial intelligence algorithms such as computational intelligence, machine learning, and mathematical programming, including: Research on intelligent optimization algorithms based on deep neural networks and time series prediction; Research on evolutionary federated learning algorithm based on knowledge transfer and privacy protection; Research on distributed zeroth-order optimization algorithm based on probability model. In recent years, we have won the second prize for excellent papers by young teachers from the Shaanxi Mathematics Association and the third prize for excellent papers by young teachers from the Shaanxi Industrial and Applied Mathematics Association. Published 15 academic papers in well-known journals in the field of artificial intelligence both domestically and internationally, with a total impact factor greater than 100, including IEEE Internet of Things Journal, IEEE Transactions on Artificial Intelligence, IEEE Transactions on Emerging Topics in Computational Intelligence, IEEE Geoscience and Remote Sensing Letters, Information Sciences, Expert System with Applications, Applied Soft Computing, Complex & Intelligent Systems, etc.
The Speech title: Neural network-assisted evolutionary algorithm for multimodal multi-objective optimization problems
Abstract: In the multimodal multi-objective optimization problems (MMOPs), the distant solutions in the decision space have similar objective values, and these solutions with good diversity provide rich choices for decision makers. Traditional evolutionary algorithms have developed various diversity maintenance mechanisms to ensure the full search of decision space. However, these mechanisms do not make full use of the distribution and variation of PSs hidden in historical data, so their adaptive ability and environmental perception ability are poor. Inspired by data-driven evolutionary algorithms, it is a very promising research to use intelligent neural network models to learn knowledge hidden in historical data and assist evolutionary algorithms to explore PSs with different distributions. Therefore, this report will introduce the research of neural network-assisted evolutionary algorithm for solving MMOPs, including the shortcomings of traditional evolutionary algorithms for solving MMOPs, evolutionary algorithm design based on predictive neural network , and the help of graph learning to the evolution process.
Feng-Feng Wei
Associate Professor
South China University of Technology, China
Bio: Feng-Feng Wei is an assistant professor at the School of Computer Science and Engineering, South China University of Technology. Her main research areas include swarm intelligence, evolutionary computing, distributed optimization, and data-driven optimization. In recent years, she has published more than 20 papers in top international journals, including IEEE Trans. Evolutionary Computation, IEEE Trans. Services Computing, et al. She is served as a reviewer for multiple international journals and conferences, including IEEE Trans. Evolutionary Computation, IEEE Trans. SMC-Systems, IEEE Trans. Neural Networks and Learning Systems, et al. She has got the first prize of the Natural Science Award of the Science and Technology Award of Guangdong Artificial Intelligence Industry Association and the Best Paper Award of the 4th International Distributed Artificial Intelligence Conference. She is hosting a national level project. Besides, she has been participating as a key member in the first batch of National Science and Technology Innovation 2030 New Generation Artificial Intelligence project, and National Key Research and Development Program Strategic Science and Technology Innovation Cooperation projects at the provincial and ministerial levels.
The Speech title: Data-Driven Evolutionary Computation in Distributed Scenarios
Abstract: With the rapid development of distributed computing paradigms such as crowdsourcing, edge computing, et al., more data are collected and processed in a distributed manner. Due to the transmission cost or data privacy, these data cannot be fused and new challenges such as partial information, non-i.i.d. data and environmental-aware noises. To solve these problems, we study data-driven evolutionary computation in distributed scenarios. Firstly, data-driven evolutionary framework for distributed optimization is studied under partial information. Secondly, consistent learning strategy is proposed for non-i.i.d. data. Finally, the cooperative coevolutionary algorithm is developed to alleviate the influence of environmental-aware noises. Experimental results show that the proposed methods have satisfactory performance for data-driven distributed optimization.
Songbai Liu
Assistant Professor
Shenzhen University, China
Bio: Songbai Liu (Member IEEE) received the Ph.D. degree from Department of Computer Sciences, City University of Hong Kong, in 2022. He is currently an assistant professor in College of Computer Science and Software Engineering, Shenzhen University. He has published over 20 papers in prestigious journals such as IEEE Transactions on Evolutionary Computation, IEEE Transactions on Cybernetics, IEEE Transactions on Systems, Man, and Cybernetics, and IEEE Transactions on Emerging Topics in Computational Intelligence. Additionally, he has served as a reviewer for these and other leading journals and conference in the field of evolutionary computation and artificial intelligence. His research interests include multiobjective optimization, evolutionary algorithms + machine learning, and their applications.
The Speech title: Data-driven Learnable Evolutionary Algorithms for Complex Multiobjective Optimization
Abstract: Evolutionary algorithms are effective for solving complex optimization problems, especially those involving multiple objectives. However, their significant computational demands and inherent stochasticity can slow down convergence to global optima, particularly in intricate scenarios. In this talk, I will discuss our efforts in accelerating complex multiobjective optimization through data-driven learnable evolutionary generators. We train a model on elaborately prepared dataset to learn how to improve solutions in various optimization contexts, including large-scale, dynamic, constrained, and expensive multiobjective optimization problems. By studying how solutions improve from suboptimal to superior, the model learns important insights into performance improvements. We then create learnable evolutionary generators based on these insights. These generators produce higher-quality offspring solutions by concurrently exploring the learned representation spaces. This approach accelerates the population’s convergence toward global optimality, steering the search process in more promising directions and enhancing overall efficiency.
Hao Hao
Postdoctoral
Shanghai Jiao Tong University, China
Bio: Dr. Hao Hao earned his Ph.D. from East China Normal University and subsequently conducted postdoctoral research at the Natural Sciences Institute of Shanghai Jiao Tong University. He has led several projects funded by the National Natural Science Foundation of China and the China Postdoctoral Science Foundation, and has been honored as a Shanghai Super Postdoctoral Fellow.
Dr. Hao serves as a reviewer for multiple prestigious international journals, such as Science China Information Sciences, IEEE Transactions on Evolutionary Computation, and Swarm and Evolutionary Computation. Additionally, he reviews for significant conferences like ECAI and AAAI. He has published over 10 academic papers, holds 2 invention patents, and has earned awards in international competitions, including NeurIPS 2021 and CVPR 2022.
Dr. Hao's research focuses on Bayesian optimization, surrogate models, and expensive optimization. By introducing innovative relation model, he has significantly enhanced the accuracy and predictive capabilities of surrogate models, offering new perspectives and tools for tackling expensive optimization challenges.
The Speech title: Expensive Optimization via Relation
Abstract: Expensive optimization problems pose significant challenges to traditional gradient-free optimization methods due to their high evaluation costs. Surrogate-assisted evolutionary algorithms (SAEAs) effectively address these challenges by using surrogate models to replace expensive evaluation functions. Designing an efficient surrogate model is crucial to the success of these algorithms. In recent years, constructing surrogate models based on the relationships between solutions has emerged as a novel modeling approach, following regression and classification models. This report proposes a general and comprehensive relational model framework and details specific implementation strategies.
Peilan Xu
Lecturer
Nanjing University of Information Science and Technology, China
Bio: Peilan Xu, Ph.D., Master's Supervisor, Lecturer. Dr. Peilan Xu graduated with a Ph.D. in Computer Science and Technology from the University of Science and Technology of China in 2022. He has conducted extensive research in the field of artificial intelligence, focusing on intelligent optimization and decision-making. His research interests include swarm intelligence, reinforcement learning, multi-agent systems, and their applications. Dr. Xu has published over 30 papers in prestigious national and international academic journals and conferences such as IEEE TEVC and ACM TELO.
The Speech title: Research on Cooperative coevolution for Large Scale Optimization
Abstract: In the era of data and intelligence, a large number of optimization problems abound in the practice of social production and life. Such problems often have complex characteristics, such as non-convex, multimodal, non-frivolous, etc., and it is even difficult to establish an accurate mathematical model. Evolutionary algorithms are widely used to solve complex optimization problems because of their swarm search characteristics. However, the proliferation of problem dimensions poses a serious challenge to the scalability of evolutionary algorithms. This report firstly introduces the cooperative coevolutionary algorithms for solving large-scale optimization problems based on the idea of divide-and-conquer, and then presents my research progress for constrained and unconstrained large-scale optimization problems.
