"Empirical Methods"

7/01/05: 10:30am-12:00pm, webcast; 12:30pm-2pm, webcast
11th Floor Large Conference Room
Host: Patrick Pantel, schedule

Abstract: This is the tutorial that Paul will give at AAAI05 on July 10. Tutorial to be held in two sessions: Session 1: 10:30am - 12:00pm and Session 2: 12:30pm - 2:00pm. The purpose of the tutorial is to introduce empirical tools including experiment and metrics design, exploratory data analysis, statistical hypothesis testing, and modeling. Through case studies and other examples, the tutorial introduces a variety of evaluation methods, including methods for visualizing, summarizing and detecting clues in data; for designing controlled experiments; and for testing hypotheses, both with classical tests and with new, computer-intensive bootstrap and randomization methods. The case studies are organized around ten lessons that are helpful when conducting evaluations (e.g., "follow the variance," and "evaluation begins with claims."). The tutorial does not require any prior knowledge of the subject, and specifically will not require background in statistics. The tutorial will therefore is suitable for the general audience

About Paul Cohen: Paul Cohen is author of the book Empirical Methods for Artificial Intelligence (MIT Press, 1995). He has published widely on methodological issues and other topics in AI.

Last updated: Mon Jun 19 17:44:06 2006