PublishScience PreprintsSYNTACTIC NEUTRALIZATION: A SEMANTIC-FUNCTIONAL PERSPECTIVESIrodaXolmurodovaauthorThis article analyzes the problem of automatic context-based disambiguation of homonymous lexemes in the Uzbek language from linguistic and methodological perspectives. The study was conducted on selected Uzbek homonymous lexemes represented in both Latin and Cyrillic scripts using two contextual models, UzBERT and XLM-R. Contexts corresponding to different meanings of each lexeme were compared using cosine similarity, and intra-class and inter-class similarity scores, as well as the separation gap between them, were calculated. The results were further validated using the internationally recognized F1 evaluation metric: UzBERT achieved an F1 score above 80% in five out of six test cases, whereas XLM-R demonstrated substantial instability, with scores ranging from 20% to 100%. These findings indicate that language-specific pretrained models provide more stable performance in distinguishing homonymy in Uzbek than multilingual models.2026-01-01Article

For work being deposited by its own author: In self-archiving this collection of files and associated bibliographic metadata, I grant PublishScience Preprints the right to store them and to make them permanently available publicly for free on-line. I declare that this material is my own intellectual property and I understand that PublishScience Preprints does not assume any responsibility if there is any breach of copyright in distributing these files or metadata. (All authors are urged to prominently assert their copyright on the title page of their work.)

For work being deposited by someone other than its author: I hereby declare that the collection of files and associated bibliographic metadata that I am archiving at PublishScience Preprints is in the public domain. If this is not the case, I accept full responsibility for any breach of copyright that distributing these files or metadata may entail.

Clicking on the Deposit Item Now button indicates your agreement to these terms.