Pareto-Based Multiobjective Machine Learning
Yaochu Jin
Professor
University of Surrey, UK
Abstract: Machine learning is inherently a multiobjective task. Traditionally, however, either only one of the objectives is adopted as the cost function or multiple objectives are aggregated to a scalar cost function. This can be mainly attributed to the fact that most conventional learning algorithms can only deal with a scalar cost function. Over the last decade, efforts on solving machine learning problems using the Pareto-based multiobjective optimization methodology have gained increasing impetus, particularly due to the great success of multiobjective optimization using evolutionary algorithms and other population-based stochastic search methods. It has been shown that Pareto-based multiobjective learning approaches are more powerful compared to learning algorithms with a scalar cost function in addressing various topics of machine learning, such as clustering, feature selection, improvement of generalization ability, knowledge extraction, and ensemble generation. One common benefit of the different multiobjective learning approaches is that a deeper insight into the learning problem
can be gained by analyzing the Pareto front composed of multiple Pareto-optimal solutions. This paper provides an overview of the existing research on multiobjective machine learning, focusing on supervised learning. In addition, a number of case studies are provided to illustrate the major benefits of the Pareto-based approach to machine learning, e.g., how to identify interpretable models andmodels that can generalize on unseen data from the obtained Pareto-optimal solutions. Three approaches to Pareto-based multiobjective ensemble generation are compared and discussed in detail. Finally, potentially interesting topics in multiobjective machine learning are suggested.
Biography: Yaochu Jin received the B.Sc., M.Sc., and Ph.D. degrees from Zhejiang University, Hangzhou, China, in 1988, 1991, and 1996 respectively, and the Dr.-Ing. degree from Ruhr University Bochum, Germany, in 2001. He is currently a Professor and Chair in Computational Intelligence, Department of Computing, University of Surrey, UK, where he heads the Nature Inspired Computing and Engineering (NICE) Group. Before joining Surrey, he was a Principal Scientist and Group Leader with the Honda Research Institute Europe in Germany. His research interests include computational approaches to a systems-level understanding evolution, learning and development in biology, and bio-inspired methods for solving complex engineering problems.
He won the Best Paper Award of the 2010 IEEE Symposium on Computational Intelligence in Bioinformatics and Computational Biology. His current research is funded by EU FP7, UK EPSRC and industries. Dr. Jin is an Associate Editor of BioSystems, the IEEE Transactions on Neural Networks and Learning Systems, IEEE Transactions on Systems, Man, and Cybernetics, Part
C: Applications and Reviews, IEEE Transactions on NanoBioscience, IEEE Computational Intelligence Magazine, and International Journal of Fuzzy Systems. He is also an Area Editor of Soft Computing. He is presently Chair of the Intelligent Systems Applications Technical Committee and an elected member of AdCom (2012-2014) of the IEEE Computational Intelligence Society.
Dr. Jin has delivered over ten invited keynote speeches on morphogenetic robotics, developmental neural systems, modeling, analysis and synthesis of gene regulatory networks, evolutionary aerodynamic design optimization and multi-objective learning at international conferences. He is a Fellow of British Computer Society and Senior Member of IEEE. See http://www.surrey.ac.uk/computing/people/yaochu_jin/index.htm for more details.