Automatic Acquisition of Ranked Qualia Structures from the Web

Proceedings of the 45th Annual Meeting of the Association of Computational Linguistics, pages 888–895, Prague, Czech Republic, June 2007. c⃝2007 Association for Computational Linguistics Automatic Acquisition of Ranked Qualia Structures from the Web1 Philipp Cimiano Inst. AIFB, University of Karlsruhe Englerstr. 11, D-76131 Karlsruhe cimiano@aifb.uni-karlsruhe.de Johanna Wenderoth Inst. AIFB, University of Karlsruhe Englerstr. 11, D-76131 Karlsruhe jowenderoth@googlemail.com Abstract This paper presents an approach for the au- tomatic acquisition of qualia structures for nouns from the Web and thus opens the pos- sibility to explore the impact of qualia struc- tures for natural language processing at a larger scale. The approach builds on ear- lier work based on the idea of matching spe- cific lexico-syntactic patterns conveying a certain semantic relation on the World Wide Web using standard search engines. In our approach, the qualia elements are actually ranked for each qualia role with respect to some measure. The specific contribution of the paper lies in the extensive analysis and quantitative comparison of different mea- sures for ranking the qualia elements. Fur- ther, for the first time, we present a quan- titative evaluation of such an approach for learning qualia structures with respect to a handcrafted gold standard. 1 Introduction Qualia structures have been originally introduced by (Pustejovsky, 1991) and are used for a variety of purposes in natural language processing (NLP), such as for the analysis of compounds (Johnston and Busa, 1996) as well as co-composition and coercion (Pustejovsky, 1991), but also for bridging reference resolution (Bos et al., 1995). Further, it has also 1The work reported in this paper has been supported by the X-Media project, funded by the European Commission under EC grant number IST-FP6-026978 as well by the SmartWeb project, funded by the German Ministry of Research. Thanks to all our colleagues for helping to evaluate the approach. been argued that qualia structures and lexical seman- tic relations in general have applications in informa- tion retrieval (Voorhees, 1994; Pustejovsky et al., 1993). One major bottleneck however is that cur- rently qualia structures need to be created by hand, which is probably also the reason why there are al- most no practical NLP systems using qualia struc- tures, but a lot of systems relying on publicly avail- able resources such as WordNet (Fellbaum, 1998) or FrameNet (Baker et al., 1998) as source of lex- ical/world knowledge. The work described in this paper addresses this issue and presents an approach to automatically learning qualia structures for nouns from the Web. The approach is inspired in recent work on using the Web to identify instances of a re- lation of interest such as in (Markert et al., 2003) and (Etzioni et al., 2005). These approaches rely on a combination of the usage of lexico-syntactic pattens conveying a certain relation of interest as described in (Hearst, 1992) with the idea of using the web as a big corpus (cf. (Kilgariff and Grefenstette, 2003)). Our approach directly builds on our previous work (Cimiano and Wenderoth, 2005) an adheres to the principled idea …

Content truncated to 500 words