English

CALE : Concept-Aligned Embeddings for Both Within-Lemma and Inter-Lemma Sense Differentiation

Computation and Language 2026-01-26 v2

Abstract

Lexical semantics is concerned with both the multiple senses a word can adopt in different contexts, and the semantic relations that exist between meanings of different words. To investigate them, Contextualized Language Models are a valuable tool that provides context-sensitive representations that can be used to investigate lexical meaning. Recent works like XL-LEXEME have leveraged the task of Word-in-Context to fine-tune them to get more semantically accurate representations, but Word-in-Context only compares occurrences of the same lemma, limiting the range of captured information. In this paper, we propose an extension, Concept Differentiation, to include inter-words scenarios. We provide a dataset for this task, derived from SemCor data. Then we fine-tune several representation models on this dataset. We call these models Concept-Aligned Embeddings (CALE). By challenging our models and other models on various lexical semantic tasks, we demonstrate that the proposed models provide efficient multi-purpose representations of lexical meaning that reach best performances in our experiments. We also show that CALE's fine-tuning brings valuable changes to the spatial organization of embeddings.

Keywords

Cite

@article{arxiv.2508.04494,
  title  = {CALE : Concept-Aligned Embeddings for Both Within-Lemma and Inter-Lemma Sense Differentiation},
  author = {Bastien Liétard and Gabriel Loiseau},
  journal= {arXiv preprint arXiv:2508.04494},
  year   = {2026}
}

Comments

Accepted at EACL 2026

R2 v1 2026-07-01T04:37:29.435Z