Berkeley Initiative in Soft Computing (BISC)
Date: Tuesday, February 5th, 2008
Time: 4-5 pm
Room: 606 Soda Hall, UC Berkeley
Title: Methods for benchmarking causal discovery
Speaker: Isabelle Guyon
Formal, widely-accepted definitions of causality have eluded philosophers of science for centuries. However the notion of causality is at the core of the scientific endeavor and also a universally accepted and intuitive notion of everyday life. From an engineering point of view, causality is a very goal-oriented notion, which can simply be defined as finding modes of action on a system, which will result in a desired outcome. For example, taking a drug to cure illness. Thus, even though causality may not find a perfect definition regrouping all the notions it encompasses in philosophy, psychology, history, law, religion, statistics, physics, and engineering, we can devise tasks, which require some notion of causality to perform successful predictions of the consequences of actions. We have stated a project to benchmark causal discovery algorithms called the "Causality Workbench" (http://clopinet.com/causality). In this presentation, we will explain how we proceed to use real data from a variety of domains to devise tasks and quantitatively evaluate the performance of algorithms.
This project is funded by the NSF Grant N0 ECCS-0725746. As part of the project we are organizing a challenge (Deadline April 30, 2008) to be discussed at a WCCI (June 2008); proceedings published in JMLR. Several prizes will be awarded.
== About the speaker: Isabelle Guyon is a researcher in machine learning and an independent consultant. Prior to starting her consulting practice in 1996, she worked at AT&T Bell Laboratories, where she pioneered applications of neural networks to pen computer interfaces and invented Support Vector Machines (in collaboration with B. Boser and V. Vapnik). Isabelle Guyon holds a Ph.D. degree in Physical Sciences of the University Pierre and Marie Curie of Paris, France. She is vice-president of the Unipen foundation, action editor of the Journal of Machine Learning Research, and competition chair of the IJCNN conference.