Year: 2026
Type: conference-paper
Abstract: SemEval 2026 task 5 asks us to provide a program to try to match the human ratings of senseappropriateness of a particular word in a series of very structured, very short stories.Our system 1 associates a fixed list of 50 words with each WordNet synset, and computes several scores for each of the phrases in the story, to determine how closely the phrase matches the wordlist.We received near-chance results, in spite of several different approaches to building and employing sets of word-lists.The stories in this dataset are designed to be ambiguous, and every story contains words associated with at least two senses of the target word.We now believe that our system's approach is inappropriate for this dataset.
Author: Stephen Taylor
Institution: University of West Bohemia in Pilsen
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: hybrid
OpenAlex ID: W7166853978