Selectional constraints: an information-theoretic model and its computational realization.

By Philip Resnik. 1996

Subfields: Natural Language Processing, Information Theory

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Natural Language Processing

In Information Theory

Citations in the corpus

Probing for semantic evidence of composition by means of simple classification tasks (2016, methodology): “Possible approaches to estimating plausibility without human annotation include using n-gram statistics on simple argument/predicate combinations (Rashkin et al., 2016) or making use of se-lectional preference modeling (Resnik, 1996; Erk, 2007; S´eaghdha, 2010).”
Assessing Composition in Sentence Vector Representations (2018, methodology): “We leave incorporation of more sophisticated selectional preference methods (Van de Cruys, 2014; Resnik, 1996) for future work.”
A Structured Vector Space Model for Word Meaning in Context (2008, methodology): “In linguistics, expectations, in the form of selectional restrictions and selectional preferences, have long been used in semantic theories (Katz and Fodor, 1964; Wilks, 1975), and more recently induced from corpora (Resnik, 1996; Brockmann and Lapata, 2003).”
A Simple, Similarity-based Model for Selectional Preferences (2007, methodology): “Some approaches have used WordNet for the generalization step (Resnik, 1996; Clark and Weir, 2001; Abe and Li, 1993), others EM-based clustering (Rooth et al., 1999).”
Redundancy and reduction: Speakers manage syntactic information density (2010, methodology): “Specifically, Resnik shows that the relative entropy of a verb v with regard to the semantic classes of its objects ci; S ¼ P i Pðci j vÞ log Pðci j vÞ PðciÞ , is correlated with the verb’s rate of object drop (Resnik, 1996, pp. 149–151).”
Toddlers encode similarities among novel words from meaningful sentences. (2015, unknown): “Cross-sentential positional similarity is of particular interest because encoding this information could help children learn semantic similarities; as noted above, words that are used in similar sentence positions often have similar semantic roles (see also Boland & Tanenhaus, 1991; Resnik, 1996).”
Predicting query performance (2002, background): “Additionally, speculation has been made about using the dispersion of the top documents as a measure of query difficulty[18], and a nearly identical mathematical framework has been used to model selectional preferences in natural language[17].”