A system developed by our researcher secured second place in the international SemEval-2026 competition focused on narrative similarity. Its developer, Marcin Sawiński, presented his research at ACL 2026 in San Diego, United States, one of the world’s leading international conferences on natural language processing.
During the conference, Marcin Sawiński presented his paper entitled “FactUEP at SemEval-2026 Task 4: Structured Narrative Similarity Scoring with Aspect Decomposition and Weak-Signal Gating”. The research explored whether large language models (LLMs) can identify similarities between the plots of two stories. The proposed system evaluates three aspects separately: the underlying theme, the sequence of events, and the outcome of each story. Cases where the model’s confidence is low are subjected to further analysis. The system achieved near-human performance and ranked second in the international SemEval-2026 narrative similarity competition (Task 4).
The full paper is available in the ACL Anthology. DOI: 10.18653/v1/2026.semeval-1.147.
ACL 2026 (the 64th Annual Meeting of the Association for Computational Linguistics) took place on July 2–7, 2026, in San Diego, California, United States. It is one of the world’s leading academic conferences dedicated to natural language processing and computational linguistics, bringing together researchers from around the globe.
The research was funded by Poznań University of Economics and Business through the subsidy provided by the Polish Ministry of Science and Higher Education and the Regional Initiative of Excellence programme (IREG project).