APPLIED TEXTUAL ENTAILMENT : A generic framework to capture shallow semantic inference

ISBN-13: 9783639151206 / Angielski / Miękka / 2009 / 132 str.

Oren Glickman

Opis

This book introduces the applied notion of textualentailment as a generic empirical task that capturesmajor semantic inferences across many applications.Textual Entailment addresses semantic inference as adirect mapping between language expressions andabstracts the common semantic inferences as neededfor text based Natural Language Processingapplications. The book defines the task and describesthe creation of a benchmark dataset for textualentailment along with proposed evaluation measures.It further describes how textual entailment can beapproximated and modeled at the lexical level andproposes a lexical reference subtask and acorrespondingly derived dataset. The book furtherproposes a general probabilistic setting that caststhe applied notion of textual entailment inprobabilistic terms. This proposed setting mayprovide a unifying framework for modeling uncertainsemantic inferences from texts. Finally, the bookpresents a novel acquisition algorithm to identifylexical entailment relations from a single corpusfocusing on the extraction of verb paraphrases.

O autorze

Oren Glickman completed his Ph.D in Computer Science at Bar-Ilan
University and holds an M.Sc in Computational Linguistics from
Carnegie Melon University and a B.Sc in Math and Computer Science
from the Hebrew University. Previously he directed research at
Shopping.com (eBay) and was a member of the research staff at
Digital Equipment Corporation.