TartuNLP @ AXOLOTL-24: Leveraging Classifier Output for New Sense Detection in Lexical Semantics

TL;DR

This work tackles diachronic word sense change detection by reframing both AXOLOTL-24 subtasks as binary matching problems between usage examples and sense definitions. It employs a GlossBERT–style cross-encoder method implemented with multilingual XLM-RoBERTa adapters to enable efficient, language-specific fine-tuning across Finnish, Russian, and German. Subtask 1 identifies novel senses by assigning the most probable existing gloss to new period usages or signaling a new sense when confidence is low, while Subtask 2 maps predicted novel senses to Wiktionary definitions to provide definitions without generating new ones. The approach is lightweight and scalable, achieving third place in Subtask 1 and first place in Subtask 2, with implications for cross-linguistic diachronic lexical semantics and reproducible NLP pipelines.

Abstract

We present our submission to the AXOLOTL-24 shared task. The shared task comprises two subtasks: identifying new senses that words gain with time (when comparing newer and older time periods) and producing the definitions for the identified new senses. We implemented a conceptually simple and computationally inexpensive solution to both subtasks. We trained adapter-based binary classification models to match glosses with usage examples and leveraged the probability output of the models to identify novel senses. The same models were used to match examples of novel sense usages with Wiktionary definitions. Our submission attained third place on the first subtask and the first place on the second subtask.