Paper: Category-Based Pseudowords

ACL ID N03-2023
Title Category-Based Pseudowords
Venue Human Language Technologies
Session Short Paper
Year 2003

A pseudoword is a composite comprised of two or more words chosen at random; the individual occurrences of the original words within a text are replaced by their conflation. Pseudowords are a useful mechanism for evaluating the im- pact of word sense ambiguity in many NLP applications. However, the standard method for constructing pseudowords has some draw- backs. Because the constituent words are cho- sen at random, the word contexts that surround pseudowords do not necessarily reflect the con- texts that real ambiguous words occur in. This in turn leads to an optimistic upper bound on algorithm performance. To address these draw- backs, we propose the use of lexical categories to create more realistic pseudowords, and eval- uate the results of different variations of this idea against th...