Detecting Compositionality of Verb-Object Combinations using Selectional Preferences

McCarthy, Diana, Venkatapathy, Sriram and Joshi, Aravind K (2007) Detecting Compositionality of Verb-Object Combinations using Selectional Preferences. In: Proceedings of the 2007 Joint Conference on Empirical Methods in Natural Language Processing and Computational Natural Language Learning (EMNLP-CoNLL), Prague, Czech Republic.

Full text not available from this repository.
Official URL: http://www.aclweb.org/

Abstract

In this paper we explore the use of selectional preferences for detecting non-compositional verb-object combinations. To characterise the arguments in a given grammatical relationship we experiment with three models of selectional preference. Two use WordNet and one uses the entries from a distributional thesaurus as classes for representation. In previous work on selectional preference acquisition, the classes used for representation are selected according to the coverage of argument tokens rather than being selected according to the coverage of argument types. In our distributional thesaurus models and one of the methods using WordNet we select classes for representing the preferences by virtue of the number of argument types that they cover, and then only tokens under these classes which are representative of the argument head data are used to estimate the probability distribution for the selectional preference model. We demonstrate a highly significant correlation between measures which use these`type-based' selectional preferences and compositionality judgements from a data set used in previous research. The type-based models perform better than the models which use tokens for selecting the classes. Furthermore, the models which use the automatically acquired thesaurus entries produced the best results. The correlation for the thesaurus models is stronger than any of the individual features used in previous research on the same dataset.

Item Type: Conference or Workshop Item (Paper)
Additional Information: Originality: This was the first use of selectional preferences for detecting compositionality of word co-occurrences. We proposed a new method for acquiring selectional preferences based on argument head word types which reduces the impact of ambiguity. Rigour: the models were evaluated on a dataset of 638 verb-object combinations. Significance: Our selectional preference models using types outperformed previous models which use tokens. Furthermore our models using automatically produced thesauruses outperformed all WordNet models that we tried and all other individual measures used previously on the same dataset. Impact: Preference models are used for information extraction and we believe that our models will be particularly useful because they are obtained from raw text (not man-made thesauruses) and so can be used for any language. Outlet: EMNLP is a prestigious international conference with acceptance rates close to 25%. Papers received 3 reviews in a double-blind review process.
Schools and Departments: School of Engineering and Informatics > Informatics
Depositing User: Diana Frances McCarthy
Date Deposited: 06 Feb 2012 19:37
Last Modified: 13 Apr 2012 08:41
URI: http://sro.sussex.ac.uk/id/eprint/21516
📧 Request an update