tailieunhanh - Báo cáo khoa học: "Integrating Pattern-based and Distributional Similarity Methods for Lexical Entailment Acquisition"

This paper addresses the problem of acquiring lexical semantic relationships, applied to the lexical entailment relation. Our main contribution is a novel conceptual integration between the two distinct acquisition paradigms for lexical relations – the patternbased and the distributional similarity approaches. The integrated method exploits mutual complementary information of the two approaches to obtain candidate relations and informative characterizing features. | Integrating Pattern-based and Distributional Similarity Methods for Lexical Entailment Acquisition Shachar Mirkin School of Computer Science and Engineering The Hebrew University Jerusalem Israel 91904 mirkin@ Abstract This paper addresses the problem of acquiring lexical semantic relationships applied to the lexical entailment relation. Our main contribution is a novel conceptual integration between the two distinct acquisition paradigms for lexical relations - the patternbased and the distributional similarity approaches. The integrated method exploits mutual complementary information of the two approaches to obtain candidate relations and informative characterizing features. Then a small size training set is used to construct a more accurate supervised classifier showing significant increase in both recall and precision over the original approaches. 1 Introduction Learning lexical semantic relationships is a fundamental task needed for most text understanding applications. Several types of lexical semantic relations were proposed as a goal for automatic acquisition. These include lexical ontological relations such as synonymy hyponymy and meronymy aiming to automate the construction of WordNet-style relations. Another common target is learning general distributional similarity between words following Harris Distributional Hypothesis Harris 1968 . Recently an applied notion of entailment between lexical items was proposed as capturing major inference needs which cut across multiple semantic relationship types see Section 2 for further background . The literature suggests two major approaches for learning lexical semantic relations distributional similarity and pattern-based. The first approach recognizes that two words or two multiword terms are semantically similar based on Ido Dagan Maayan Geffet Department of Computer Science Bar-Ilan University Ramat Gan Israel 52900 dagan zitima @ distributional similarity of the different contexts .