Tutorial T4: Meta-Learning & Algorithm Selection
Monday, August 18, 11:00–12:30
Pavel Brazdil, Carlos Soares and Joaquin Vanschoren
Copies of slide are available from this page.You can download the zip file containing, 2 groups of slides and open it with a password. You can obtain the password from the presenters. In your request, please declare that you have registered for the ECAI-2014 tutorial.
Algorithm Selection and configuration are increasingly relevant today. Researchers and practitioners from all branches of science and technology face a large choice of parameterized algorithms, with little guidance as to which techniques to use. Moreover, data mining challenges frequently remind us that algorithm selection and configuration are crucial in order to achieve the best performance and drive industrial applications.
Meta-learning leverages knowledge of past algorithm applications to select the best techniques for future applications, and offers effective techniques that are superior to humans both in terms of the end result and especially in a limited time.
In this tutorial, we elucidate the nature of algorithm selection and how it arises in many diverse domains, such as machine learning, data mining, optimization and SAT solving. We show that it is possible to use meta-learning techniques to identify the potentially best algorithm(s) for a new task, based on meta-level information and prior experiments. We also discuss the prerequisites for effective meta-learning systems, and how recent infrastructures, such as OpenML.org, allow us to build systems that effectively advice users on which algorithms to apply.
The intended audience includes researchers (Ph.D.’s), research students and practitioners interested to learn about, or consolidate their knowledge about:
- the state-of-the-art in algorithm selection and algorithm configuration,
- how to use Data Mining software and platforms to select algorithms in practice,
- how to provide advice to end users about which algorithms to select in diverse domains, including optimization, SAT etc. and incorporate this knowledge in new platforms.
The participants should bring their own laptops.
The prospective participants of the tutorial could glance at the objectives of the related ECAI workshop MetaSel - Meta-learning & Algorithm Selection.
Short bio of the presenter(s)
Pavel Brazdil (pbrazdil AT inescporto.pt) obtained his Ph.D. in AI at the Univ. Edinburgh in 1981 and since then has been lecturing at the Univ. Porto, Portugal (since 1998 as Full Professor). He is a member of LIAAD-INESC Tec Laboratory. His interests lie in data mining, machine learning, meta-learning and text mining. He has supervised 11 Ph.D. students, published/edited 7 books, various chapters in books, journals articles and 40+ articles listed on ISI/DBLP which have achieved 2000+ citations on Google Scholar. He has chaired/organized 10 conferences/workshops (e.g. ECML-91, ECML-93, ECML/PKDD 2005 etc.). He is a Fellow of ECCAI.
Carlos Soares (csoares AT feup.up.pt) received a Ph.D. in Computer Science from U.Porto. He works as Associate Professor at the Faculty of Engineering of U.Porto. He is a researcher at INESC TEC, focusing on Data Mining and Business Intelligence. He has participated in 20+ R&ID projects, published/edited several books and 40+ papers in journals and conferences proceedings indexed by ISI. He has participated in the organization of e.g., KDD-09 and ECML-PKDD-12, and he will be Programme Chair for ECML-PKDD-15. He was awarded the Scientific Merit and Excellence Award of the Portuguese AI Association.
Joaquin Vanschoren (j.vanschoren AT tue.nl) is assistant professor at the Eindhoven University of Technology (TU/e). He runs OpenML.org, an open science platform for machine learning and meta-learning integrated in WEKA, R, and other popular machine learning environments. His research interests include meta-learning, web-scale machine learning, and data science. He obtained several demo and application awards and has been invited speaker on several conferences. He co-organized ECMLPKDD 2013, BeneLearn 2010-2011, as well as the "Silver" (ECMLPKDD) and "Planning to learn" (ECAI) workshops.
Reference person: Pavel Brazdil