Textual Knowledge Representation and Exploration

Abstract

How can we capture the knowledge expressed in large amounts of text? And how can we allow people to easily explore it, when gaining knowledge about a topic? Common knowledge representation paradigms encode knowledge in a formal language, whose vocabulary is typically pre-specified and hence is inherently limited in scope. In this talk I will outline a research direction that aims to encode textual knowledge based on the available natural language vocabulary and structure. First, we propose identifying the set of individual propositions expressed in sentences, and representing them in a simple canonical language-based structure. Then, we propose consolidating these propositions and inducing a global structure over them based on relevant semantic relationships. In particular, we focus initially on identifying equivalent propositions and how some pieces of information elaborate over others. I will review few research activities along these lines, including a proposition extraction tool, context-sensitive lexical inference and the extraction of lexical inferences from structured knowledge resources, and will illustrate the potential of this approach for text exploration.

Biography:

Ido Dagan is an Associate Professor at the Department of Computer Science of Bar-Ilan University, Israel. His interests are in applied semantic processing, focusing on the development of generic textual inference models, knowledge acquisition methods and novel application schemes that are based on them. Dagan and colleagues defined the textual entailment recognition task and organized the series of Recognizing Textual Entailment Challenges. He was the President of the Association for Computational Linguistics (ACL) in 2010 and served on its Executive Committee during 2008-2011. In that capacity, he led the establishment of the Transactions of the Association for Computational Linguistics journal. Dagan received his B.A. and Ph.D. (1992) in Computer Science from the Technion. He was a research fellow at the IBM Haifa Scientific Center (1991) and a Member of Technical Staff at AT&T Bell Laboratories (1992-1994). During 1998-2003 he was co-founder and CTO of FocusEngine and VP of Technology of LingoMotors.