ACL Findings2024
Definition generation for lexical semantic change detection
Mariia Fedorova, Andrey Kutuzov, Yves Scherrer
Abstract
We use contextualized word definitions generated by large language models as semantic representations in the task of diachronic lexical semantic change detection (LSCD). In short, generated definitions are used as 'senses', and the change score of a target word is retrieved by comparing their distributions in two time periods under comparison. On the material of five datasets and three languages, we show that generated definitions are indeed specific and general enough to convey a signal sufficient to rank sets of words by the degree of their semantic change over time. Our approach is on par with or outperforms prior non-supervised sensebased LSCD methods. At the same time, it preserves interpretability and allows to inspect the reasons behind a specific shift in terms of discrete definitions-as-senses. This is another step in the direction of explainable semantic change modeling.
Ask Lune
Know this paper in minutes, not an afternoon.
Lune can read every page of it and the top-tier papers that cite it, and gives you its claims, its evidence and its limits, each one cited to the paper it comes from. No skimming, no AI slop.
2 top-tier papers have cited it since. Know what they found before your reviewers do.
Lune can draw on
the full paper · the top-tier papers that cite it · related work
Free to start. No credit card required.
Lineage
- Definition Generation for Word Meaning Modeling: Monolingual, Multilingual, and Cross-Lingual Perspectives cites Interpretable Word Sense Representations via Definition Generation: The Case of Semantic Change Analysis
- Definition Generation for Word Meaning Modeling: Monolingual, Multilingual, and Cross-Lingual Perspectives cites Automatically Generated Definitions and their utility for Modeling Word Meaning
- Interpretable Word Sense Representations via Definition Generation: The Case of Semantic Change Analysis cites DWUG: A large Resource of Diachronic Word Usage Graphs in Four Languages
- Automatically Generated Definitions and their utility for Modeling Word Meaning cites Interpretable Word Sense Representations via Definition Generation: The Case of Semantic Change Analysis
Cited by top-tier papers2
- Automatically Generated Definitions and their utility for Modeling Word MeaningFrancesco Periti, David Alfter, Nina TahmasebiEMNLP 2024 · 2 citations
- Definition Generation for Word Meaning Modeling: Monolingual, Multilingual, and Cross-Lingual PerspectivesFrancesco Periti, Roksana Goworek, Haim Dubossarsky, Nina TahmasebiEMNLP 2025
Builds on4
- Crosslingual Generalization through Multitask FinetuningNiklas Muennighoff, Thomas Wang, Lintang Sutawika, Adam Roberts et al.ACL 2023 · 319 citations
- Interpretable Word Sense Representations via Definition Generation: The Case of Semantic Change AnalysisMario Giulianelli, Iris Luden, Raquel Fernández, Andrey KutuzovACL 2023 · 9 citations
- A Word Sense Distribution-based approach for Semantic Change PredictionXiaohang Tang, Yi Zhou, Taichi Aida, Procheta Sen et al.EMNLP Findings 2023 · 2 citations
- DWUG: A large Resource of Diachronic Word Usage Graphs in Four LanguagesDominik Schlechtweg, Nina Tahmasebi, Simon Hengchen, Haim Dubossarsky et al.EMNLP 2021 · 1 citation
Related papers
- Explaining novel senses using definition generation with open language modelsMariia Fedorova, Andrey Kutuzov, Francesco Periti, Yves ScherrerEMNLP Findings 2025
- Analyzing Semantic Change through Lexical ReplacementsFrancesco Periti, Pierluigi Cassotti, Haim Dubossarsky, Nina TahmasebiACL 2024
- Analysing Lexical Semantic Change with Contextualised Word RepresentationsMario Giulianelli, Marco Del Tredici, Raquel FernándezACL 2020 · 118 citations
- Fake it Till You Make it: Self-Supervised Semantic Shifts for Monolingual Word Embedding TasksMaurício Gruppi, Pin-Yu Chen, Sibel AdaliAAAI 2021 · 7 citations
- Swap and Predict - Predicting the Semantic Changes in Words across Corpora by Context SwappingTaichi Aida, Danushka BollegalaEMNLP Findings 2023 · 2 citations
