UCB: System Description for SemEval Task #4

Proceedings of the 4th International Workshop on Semantic Evaluations (SemEval-2007), pages 366–369, Prague, June 2007. c⃝2007 Association for Computational Linguistics UCB: System Description for SemEval Task #4 Preslav I. Nakov EECS, CS division University of California at Berkeley Berkeley, CA 94720 nakov@cs.berkeley.edu Marti A. Hearst School of Information University of California at Berkeley Berkeley, CA 94720 hearst@ischool.berkeley.edu Abstract The UC Berkeley team participated in the SemEval 2007 Task #4, with an approach that leverages the vast size of the Web in or- der to build lexically-specific features. The idea is to determine which verbs, preposi- tions, and conjunctions are used in sentences containing a target word pair, and to com- pare those to features extracted for other word pairs in order to determine which are most similar. By combining these Web fea- tures with words from the sentence context, our team was able to achieve the best results for systems of category C and third best for systems of category A. 1 Introduction Semantic relation classification is an important but understudied language problem arising in many NLP applications, including question answering, in- formation retrieval, machine translation, word sense disambiguation, information extraction, etc. This year’s SemEval (previously SensEval) competition has included a task targeting the important special case of Classification of Semantic Relations between Nominals. In the present paper we describe the UCB system which took part in that competition. The SemEval dataset contains a total of 7 se- mantic relations (not exhaustive and possibly over- lapping), with 140 training and about 70 testing sentences per relation. Sentence classes are ap- proximately 50% negative and 50% positive (“near misses”). Table 1 lists the 7 relations together with some examples. # Relation Name Examples 1 Cause-Effect hormone-growth, laugh-wrinkle 2 Instrument-Agency laser-printer, ax-murderer 3 Product-Producer honey-bee, philosopher-theory 4 Origin-Entity grain-alcohol, desert-storm 5 Theme-Tool work-force, copyright-law 6 Part-Whole leg-table, door-car 7 Content-Container apple-basket, plane-cargo Table 1: SemEval dataset: Relations with examples (context sentences are not shown). Each example consists of a sentence, two nomi- nals to be judged on whether they are in the target semantic relation, manually annotated WordNet 3.0 sense keys for these nominals, and the Web query used to obtain that example: "Among the contents of the <e1>vessel</e1> were a set of carpenters <e2>tools</e2>, several large storage jars, ceramic utensils, ropes and remnants of food, as well as a heavy load of ballast stones." WordNet(e1) = "vessel%1:06:00::", WordNet(e2) = "tool%1:06:00::", Content-Container(e2, e1) = "true", Query = "contents of the * were a" 2 Related Work Lauer (1995) proposes that eight prepositions are enough to characterize the relation between nouns in a noun-noun compound: of, for, in, at, on, from, with or about. Lapata and Keller (2005) improve on his results by using Web statistics. Rosario et al. (2002) use a “descent of hierarchy”, which charac- terizes the relation based on the semantic category of the two nouns. Girju et al. (2005) apply SVM, deci- sion trees, semantic scattering and iterative seman- 366 tic specialization, using WordNet, word sense dis- ambiguation, and …

Content truncated to 500 words