Hello. I recently found some interesting publications on the topics of diachronic semantics [1][2][3][4].
Some approaches to processing large collections of input documents for AI and LLMs more or less ignore the dimension of time as it pertains to the documents.
Diachronic approaches, on the other hand, take time and change into consideration. The meaning of words, e.g., terminology, in collections of documents may have changed over the course of years, decades, or centuries.
Thank you. I hope that these topics are also of some interest to you.
Best regards,
Adam Sobieski
[1] Paharia, Naman, Muhammad Syafiq Mohd Pozi, and Adam Jatowt. “Change Summarization of Diachronic Scholarly Paper Collections by Semantic Evolution Analysis.” In 2021 ACM/IEEE Joint Conference on Digital Libraries (JCDL), pp. 234-237. IEEE, 2021.
[2] Tahmasebi, Nina, Lars Borin, and Adam Jatowt. “Survey of Computational Approaches to Lexical Semantic Change Detection.” Computational approaches to semantic change 6, no. 1 (2021).
[3] Kutuzov, Andrey, Lilja Øvrelid, Terrence Szymanski, and Erik Velldal. “Diachronic Word Embeddings and Semantic Shifts: A Survey.” arXiv preprint arXiv:1806.03537 (2018).
[4] Wang, Jiexin, Adam Jatowt, Masatoshi Yoshikawa, and Yi Cai. “BiTimeBERT: Extending Pre-trained Language Representations with Bi-temporal Information.” In Proceedings of the 46th International ACM SIGIR Conference on Research and Development in Information Retrieval, pp. 812-821. 2023.