Universidad Politécnica de Madrid, Spain
The last few decades have seen an increase in the number of metaheuristics inspired by different natural and social phenomena. When the metaphor is stripped away, are these new algorithms any different from well-established metaheuristics? I argue that there is a lack of tools for understanding the dynamics of metaheuristics and the global structure of fitness landscapes. This talk aims to fill this gap. We will go over two network-based models of search and optimisation: local optima networks (LONs) and search trajectory networks (STNs). While LONs can be seen as compressed models of fitness landscapes, STNs emphasize the search dynamics. To cover the wide range of EvoStar interests, we will show case-studies in classical combinatorial and continuous optimisation as well as in hyper-parameters optimisation, genetic improvement and neuroevolution. Both LONs and STNs allow us to visualize realistic search spaces in ways not previously possible, and bring a whole new set of quantitative network metrics for characterizing and understanding computational search. With an emphasis on visualization, we will highlight the surprising insights these modeling tools can bring to our field.
Gabriela Ochoa is a Professor of Computing Science at the University of Stirling in Scotland. Her research lies in the foundations and applications of evolutionary algorithms and metaheuristics, with emphasis on autonomous search, fitness landscape analysis and visualization. She holds a PhD from the University of Sussex, UK, and has held academic and research positions at the University Simon Bolivar, Venezuela, and the University of Nottingham, UK. Her recent work on network-based models of fitness landscapes has enhanced their descriptive and visualization capabilities, producing a number of publications including several best-paper awards nominations. She collaborates cross-disciplines in the use of evolutionary algorithms in healthcare and conservation. She has been active in organization and editorial roles within leading Evolutionary Computation conferences such as EvoStar, GECCO and PPSN, and Journals such as Evolutionary Computation and ACM Transactions on Evolutionary Learning and Optimisation (TELO). She is a member of the executive boards of SPECIES and ACM SIGEVO and was recognised in 2020 in EvoStar for her outstanding contributions to the field.
Estimation of distribution algorithms (EDAs) are heuristic methods for evolutionary computation. In them, at each iteration, a probabilistic graphical model is induced from the selected individuals, to subsequently obtain the new population via simulation of the model. The talk will review the specialized literature in which EDAs have been applied to solve various problems in machine learning: preprocessing, feature subset selection, supervised classification, clustering, and reinforcemente learning. Different problems in machine learning that constitute challenges for the application of EDAs will be shown.
Pedro Larrañaga is Full Professor in Computer Science and Artiﬁcial Intelligence at the Universidad Politécnica de Madrid . He received the MSc degree in Mathematics (Statistics) from the University of Valladolid and the PhD degree in Computer Science from the University of the Basque Country (Excellence Award). His research interests are primarily in the areas of probabilistic graphical models, metaheuristics for optimization, machine learning classification models, and real applications, like biomedicine, bioinformatics, neuroscience, industry 4.0 and sports. He has published more than 200 papers in high impact factor journals and has supervised more than 30 PhD theses. He is fellow of the European Association for Artificial Intelligence since 2012 and fellow of the Academia Europaea and of the Asia-Pacific Artificial Intelligence Association since 2018 and 2021 respectively. He has been awarded the 2013 Spanish National Prize in Computer Science, the prize of the Spanish Association for Artificial Intelligence in 2018 and the Amity Research Award in Machine Learning in New Delhi, in 2020.