UNT-Yahoo: SuperSenseLearner: Combining SenseLearner with SuperSense and other Coarse Semantic Features

Proceedings of the 4th International Workshop on Semantic Evaluations (SemEval-2007), pages 406–409, Prague, June 2007. c⃝2007 Association for Computational Linguistics UNT-Yahoo: SuperSenseLearner: Combining SenseLearner with SuperSense and other Coarse Semantic Features Rada Mihalcea and Andras Csomai University of North Texas rada@cs.unt.edu,csomaia@unt.edu Massimiliano Ciaramita Yahoo! Research Barcelona massi@yahoo-inc.com Abstract We describe the SUPERSENSELEARNER system that participated in the English all- words disambiguation task. The system re- lies on automatically-learned semantic mod- els using collocational features coupled with features extracted from the annotations of coarse-grained semantic categories gener- ated by an HMM tagger. 1 Introduction The task of word sense disambiguation consists of assigning the most appropriate meaning to a poly- semous word within a given context. Applications such as machine translation, knowledge acquisition, common sense reasoning, and others, require knowl- edge about word meanings, and word sense disam- biguation is considered essential for all these tasks. Most of the efforts in solving this problem were concentrated so far toward targeted supervised learning, where each sense tagged occurrence of a particular word is transformed into a feature vector, which is then used in an automatic learning process. The applicability of such supervised algorithms is however limited only to those few words for which sense tagged data is available, and their accuracy is strongly connected to the amount of labeled data available at hand. Instead, methods that address all words in unre- stricted text have received significantly less atten- tion. While the performance of such methods is usu- ally exceeded by their supervised lexical-sample al- ternatives, they have however the advantage of pro- viding larger coverage. In this paper, we describe SUPERSENSE- LEARNER – a system for solving the semantic am- biguity of all words in unrestricted text. SUPER- SENSELEARNER brings together under one system the features previously used in the SENSELEARNER (Mihalcea and Csomai, 2005) and the SUPERSENSE (Ciaramita and Altun, 2006) all-words word sense disambiguation systems. The system is using a rel- atively small pre-existing sense-annotated data set for training purposes, and it learns global semantic models for general word categories. 2 Learning for All-Words Word Sense Disambiguation Our goal is to use as little annotated data as possi- ble, and at the same time make the algorithm gen- eral enough to be able to disambiguate as many content words as possible in a text, and efficient enough so that large amounts of text can be anno- tated in real time. SUPERSENSELEARNER is at- tempting to learn general semantic models for var- ious word categories, starting with a relatively small sense-annotated corpus. We base our experiments on SemCor (Miller et al., 1993), a balanced, se- mantically annotated dataset, with all content words manually tagged by trained lexicographers. The input to the disambiguation algorithm con- sists of raw text. The output is a text with word meaning annotations for all open-class words. The algorithm starts with a preprocessing stage, where the text is tokenized and annotated with part- 406 of-speech tags; collocations are identified using a sliding window approach, where a collocation …

Content truncated to 500 words