Meaning Change Detection

dc.contributor.author	Pastukhova, Iryna
dc.date.accessioned	2024-08-23T09:07:00Z
dc.date.available	2024-08-23T09:07:00Z
dc.date.issued	2024
dc.identifier.citation	Pastukhova Iryna. Meaning Change Detection. Ukrainian Catholic University, Faculty of Applied Sciences, Department of Computer Sciences. Lviv 2024, 35 p.	uk
dc.identifier.uri	https://er.ucu.edu.ua/handle/1/4674
dc.language.iso	en	uk
dc.subject	Detecting changes	uk
dc.subject	natural language processing	uk
dc.title	Meaning Change Detection	uk
dc.type	Preprint	uk
dc.status	Публікується вперше	uk
dc.description.abstracten	Detecting changes in the meaning of text after paraphrasing or editing is a chal- lenging and non-trivial task in natural language processing (NLP). It is implicitly involved in other tasks such as translation, summarisation, and style transfer. Ap- proaches to meaning change detection (or paraphrase identification) have evolved as the field of NLP has developed. Today, deep learning BERT-based models and Large Language Models (LLMs) provide state-of-the-art results. However, these methods need more interpretability and control and are computationally expensive. There are alternative methods based on linguistic and mathematical ideas that can overcome the shortcomings of LLMs and DL methods or complement them. We aim to investigate the possibilities and limitations of one such alternative ap- proach compared to state-of-the-art solutions for the paraphrase identification task.	uk