UIUC: A Knowledge-rich Approach to Identifying Semantic Relations between Nominals

Proceedings of the 4th International Workshop on Semantic Evaluations (SemEval-2007), pages 386–389, Prague, June 2007. c⃝2007 Association for Computational Linguistics UIUC: A Knowledge-rich Approach to Identifying Semantic Relations between Nominals Brandon Beamer,1,4 Suma Bhat,2,4 Brant Chee,3,4 Andrew Fister,1,4 Alla Rozovskaya,1,4 Roxana Girju1,4 Department of Linguistics1, Department of Electrical and Computer Engineering2, Department of Library and Information Science3, Beckman Institute4, University of Illinois at Urbana-Champaign {bbeamer, spbhat2, chee, afister2, rozovska, girju}@uiuc.edu Abstract This paper describes a supervised, knowledge-intensive approach to the auto- matic identification of semantic relations between nominals in English sentences. The system employs different sets of new and previously used lexical, syntactic, and semantic features extracted from various knowledge sources. At SemEval 2007 the system achieved an F-measure of 72.4% and an accuracy of 76.3%. 1 Introduction The SemEval 2007 task on Semantic Relations be- tween Nominals is to identify the underlying se- mantic relation between two nouns in the context of a sentence. The dataset provided consists of a definition file and 140 training and about 70 test sentences for each of the seven relations consid- ered: Cause-Effect, Instrument-Agency, Product- Producer, Origin-Entity, Theme-Tool, Part-Whole, and Content-Container. The task is defined as a binary classification problem. Thus, given a pair of nouns and their sentential context, the classifier decides whether the nouns are linked by the target semantic relation. In each training and test exam- ple sentence, the nouns are identified and manu- ally labeled with their corresponding WordNet 3.0 senses. Moreover, each example is accompanied by the heuristic pattern (query) the annotators used to extract the sentence from the web and the position of the arguments in the relation. (1) 041 ”He derives great joy and <e1>happiness</e1> from <e2>cycling</e2>.” WordNet(e1) = ”happiness%1:12:00::”, WordNet(e2) = ”cy- cling%1:04:00::”, Cause-Effect(e2,e1) = ”true”, Query = ”happiness from *” Based on the information employed, systems can be classified in four types of classes: (A) systems that use neither the given WordNet synsets nor the queries, (B) systems that use only WordNet senses, (C) systems that use only the queries, and (D) sys- tems that use both. In this paper we present a type-B system that re- lies on various sets of new and previously used lin- guistic features employed in a supervised learning model. 2 Classification of Semantic Relations Semantic relations between nominals can be en- coded by different syntactic constructions. We extend here over previous work that has focused mainly on noun compounds and other noun phrases, and noun–verb–noun constructions. We selected a list of 18 lexico-syntactic and se- mantic features split here into three sets: feature set #1 (core features), feature set #2 (context features), and the feature set #3 (special features). Table 1 shows all three sets of features along with their defi- nitions; a detailed description is presented next. For some features, we list previous works where they proved useful. While features F1 – F4 were selected from our previous experiments, all the other features are entirely the contribution of this research. Feature set #1: Core features This set …

Content truncated to 500 words