Colloquium Details
Learning first order logic from queries
| Author: | Marta Arias Center for Computational Learning Systems, Columbia University |
|---|---|
| Date: | March 17, 2005 |
| Time: | 3:30 |
| Location: | 220 Deschutes |
| Host: | Dejing Dou |
Abstract
In this talk I will give an overview of my work on learning first order representations from queries. Upper bounds are obtained through a learning algorithm that is able to learn an interesting subset of conjunctions of first order Horn clauses. I will highlight the difficulties of learning this class and how the algorithm solves them. Then I will introduce a machine learning system that is based on our learning algorithm, and show some results obtained in various domains, including the prediction of the mutagenicity of given molecules, prediction of prepositional phrase attachments in english sentences, and the detection of illegal positions in the game of chess.
Biography:
Dr. Arias is an Associate Research Scientist at the Center for Computational Learning Systems at Columbia University. Her interests are in machine learning and its applications to computational biology. In May 2004 she received her PhD from Tufts University for her work on quantifying the complexity of learning first order logic formulas. She received a B.S. degree in Computer Engineering from the Polytechnic University of Catalunya, Spain.