The keynotes will discuss the latest advancements on the interplay between learning and optimization.

Abstract

The more we use AI tools, the more we need to understand how they work, how they might fail, and how we can improve them. To that end, we propose a simple "DUO" model of data mining and optimziation: data mining chops up some space into regions and optimizers paint arrows between regions saying "please, go this way" or "danger, don't do that". In this combined approach, data mining and optimizers are two parts of a greater whole that acts like an agent leaning over the shoulder of an analyst that advises "ask this question next" or "ignore that problem, it is not relevant to your goals". Such agents can help us build "better" predictive models, where "better" can be either greater predictive accuracy or faster modeling time (which, in turn, enables the exploration of a wider range of options).

This talk presents a tutorial on this DUO approach and lists examples of DUO-in-action, from software engineering domains.

Biography

Tim Menzies (IEEE Fellow, Ph.D., UNSW, 1995) is a full Professor in CS at North Carolina State University where he teaches software engineering, automated software engineering, and foundations of software science. He is the directory of the RAISE lab (real world AI for SE). that explores SE, data mining, AI, search-based SE, and open access science.

He is the author of over 280 referred publications and editor of three recent books summarized the state of the art in software analytics. In his career, he has been a lead researcher on projects for NSF, NIJ, DoD, NASA, USDA (funding totalling over 12 million dollars) as well as joint research work with private companies. For 2002 to 2004, he was the software engineering research chair at NASA's software Independent Verification and Validation Facility.

Prof. Menzies is the co-founder of the PROMISE conference series devoted to reproducible experiments in software engineering (http://tiny.cc/seacraft). He is an associate editor of IEEE Transactions on Software Engineering, Communications of the ACM, ACM Transactions on Software Engineering Methodologies, Empirical Software Engineering, the Automated Software Engineering Journal the Big Data Journal, Information Software Technology, IEEE Software, and the Software Quality Journal. In 2015, he served as co-chair for the ICSE'15 NIER track. He has served as co-general chair of ICSME'16 and co-PC-chair of SSBSE'17, and ASE'12. For more, see his vita (http://menzies.us/pdf/MenziesCV.pdf or his list of publications http://tiny.cc/timpubs) or his home page http://menzies.us.

Abstract

Federated learning is a powerful machine learning paradigm for privacy-preserving machine learning. In this talk, I am going to introduce a multi-objective approach to enhancing the performance of federated learning in terms of accuracy, communication efficiency and computational complexity. I’ll start with a brief introduction to evolutionary multi-objective machine learning, followed by a presentation of two evolutionary multi-objective federated learning algorithms for optimizing the architecture of neural network models in federated learning, one for offline, and the other for real-time purposes. Finally, remaining research challenges will be outlined.

Biography

Prof. 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 Distinguished Chair, Professor in Computational Intelligence, Department of Computer Science, University of Surrey, Guildford, U.K., where he heads the Nature Inspired Computing and Engineering (NICE) Group.

He was a “Finland Distinguished Professor” of University of Jyvaskyla, Finland, and “Changjiang Distinguished Visiting Professor”, Northeastern University, China. In 2021, he was awarded the Alexander von Humboldt Professorship for Artificial Intelligence by the Federal Ministry of Education and Research, German. His main research interests include data-driven evolutionary optimization, evolutionary learning, trustworthy machine learning, and morphogenetic self-organizing systems.

Prof. Jin is presently the Editor-in-Chief of the IEEE TRANSACTIONS ON COGNITIVE AND DEVELOPMENTAL SYSTEMS and the Editor-in-Chief of Complex & Intelligent Systems. He is the recipient of the 2018 and 2021 IEEE Transactions on Evolutionary Computation Outstanding Paper Award, and the 2015, 2017, and 2020 IEEE Computational Intelligence Magazine Outstanding Paper Award. He was named by the Web of Science as “a Global Highly Cited Researcher” in 2019 and 2020. He is a Member of Academia Europaea and Fellow of IEEE.

abstract

The role of data-driven AI in Engineering and particularly in Engineering Design has made significant progress in the last years. In the first part of my presentation, I will outline the CAE/AI enhanced approach to engineering design from an industrial perspective. This will include a brief description of the main components for engineering design optimization and some concrete application examples. Topology optimization including the combination of static and crash loads will be discussed as well as a brief overview of remaining challenges for CAE/AI systems in engineering design.

In the second part of my presentation, I will introduce approaches to go beyond the tool-based AI in the engineering design process chain and enable the AI methods to improve their performance over time. Experience-based Computation: learning to optimise is an EU Horizon 2020 project that addresses the issue on how optimization can be improved through learning just like the engineer becomes more and more experienced over time. I will look at one approach inspired from data mining and introduce learned representations for engineering design. Transfer learning and the advantage of multi-task optimization will be discussed.

AI as a cooperative partner in the engineering design process will be the subject of the last part of my presentation. I will briefly introduce the general concept of cooperative intelligence and then outline some of the challenges in understanding the engineer for optimal support. Many if not most engineering design decisions are made in a team, therefore, it is necessary to go beyond the cooperative interaction between the engineer and AI as a partner, but to also study the effect that an AI system can have on the decision dynamics in a team. For this I will present first results, how AI recommendations for a very simplified task can influence the human decision making process.

The presentation will conclude with a summary and some additional issues that have to be addressed to evolve AI from a tool to a partner in Engineering and in Engineering Design.

Biography

Bernhard Sendhoff obtained a diploma in Theoretical Physics (Dipl.Phys.), in 1993, from the Ruhr-Universität Bochum, Germany, and in 1998, a PhD in Applied Physics (Artificial Intelligence). After working at Honda R&D Europe (Deutschland) GmbH, from 1999 to 2003, he has been working at the Honda Research Institute Europe GmbH from 2003 to 2011 as Chief Technology Officer and from 2011 to 2018 as the President. From 2017-2021 he was Head of Global Operations of the Honda Research Institutes and from 2019-2021 President of the Honda Research Institute Japan Co., Ltd. Japan. From 2007-2020 he was Honorary Professor of the School of Computer Science of the University of Birmingham, Great Britain. Since 2017, Bernhard Sendhoff is Operating Officer at Honda R&D Co., Ltd. Since 2021, he is Chief Executive Officer of the Global Network Honda Research Institutes and since 2008, he is Honorary Professor at the Technical University of Darmstadt, Germany. Bernhard Sendhoff is a Senior Member of the IEEE and the ACM, and a Member of the SAE. He has authored or co-authored more than 180 peer reviewed journal and conference papers and over 40 patents.

abstract

Coevolution is an old but very interesting research topic in evolutionary computation. This talk presents some of the applications of coevolution in learning and optimisation. First, we look at a classical coevolutionary learning scenario when no training data are available. In fact, no teacher information is available either. Then we examine how coevolution could be used to tackle large-scale global optimisation in the black box optimisation setting. Finally, we explore how coevolution could be harnessed to design general solvers automatically for hard combinatorial optimisation problems.

Biography

Xin Yao is a Chair Professor of Computer Science at the Southern University of Science and Technology (SUSTech), Shenzhen, China, and a part-time Professor of Computer Science at the University of Birmingham, UK. He is an IEEE Fellow and was a Distinguished Lecturer of the IEEE Computational Intelligence Society (CIS). He served as the President (2014-15) of IEEE CIS and the Editor-in-Chief (2003-08) of IEEE Transactions on Evolutionary Computation. His major research interests include evolutionary computation, ensemble learning, and their applications to software engineering. His research work won the 2001 IEEE Donald G. Fink Prize Paper Award; 2010, 2016 and 2017 IEEE Transactions on Evolutionary Computation Outstanding Paper Awards; 2011 IEEE Transactions on Neural Networks Outstanding Paper Award; and many other best paper awards at conferences. He received a Royal Society Wolfson Research Merit Award in 2012, the IEEE CIS Evolutionary Computation Pioneer Award in 2013 and the 2020 IEEE Frank Rosenblatt Award.