UWB at SemEval-2026 Task 5: Synsets and their contexts

AbstractSemEval 2026 task 5 asks us to provide a pro-gram to try to match the human ratings of sense-appropriateness of a particular word in a seriesof very structured, very short stories.Our system1 associates a fixed list of 50 wordswith each WordNet synset, and computes sev-eral scores for each of the phrases in the story,to determine how closely the phrase matchesthe wordlist.We received near-chance results, in spite ofseveral different approaches to building andemploying sets of word-lists. The stories inthis dataset are designed to be ambiguous, andevery story contains words associated with atleast two senses of the target word. We nowbelieve that our system’s approach is inappro-priate for this dataset.