← 研踪 PaperTrailarXiv:2509.11513 · v2
Unsupervised Candidate Ranking for Lexical Substitution via Holistic Sentence Semantics
首次提交 2025-09-15 09:57
版本更新 2026-11-11 21:19
雷达观测 2026-11-13 14:16
摘要
A key subtask in lexical substitution is ranking the given candidate words. A common approach is to replace the target word with a candidate in the original sentence and feed the modified sentence into a model to capture semantic differences before and after substitution. However, effectively modeling the bidirectional influence of candidate substitution on both the target word and its context remains challenging. Existing methods often focus solely on semantic changes at the target position or rely on parameter tuning over multiple evaluation metrics, making it difficult to accurately characterize semantic variation. To address this, we investigate two approaches: one based on attention weights and another leveraging the more interpretable integrated gradients method, both designed to measure the influence of context tokens on the target token and to rank candidates by incorporating semantic similarity between the original and substituted sentences. Experiments on the LS07 and SWORDS datasets demonstrate that both approaches improve ranking performance.
作者备注：A paper I recently submitted has been accepted; it shares the same title but the content is slightly different.
观测到的版本
相邻记录可能跳过中间版本；首次发现 v2 及以上也会标为版本更新。
v2 · 2026-11-11 21:19 · 最新观测版本
Unsupervised Candidate Ranking for Lexical Substitution via Holistic Sentence Semantics
A key subtask in lexical substitution is ranking the given candidate words. A common approach is to replace the target word with a candidate in the original sentence and feed the modified sentence into a model to capture semantic differences before and after substitution. However, effectively modeling the bidirectional influence of candidate substitution on both the target word and its context remains challenging. Existing methods often focus solely on semantic changes at the target position or rely on parameter tuning over multiple evaluation metrics, making it difficult to accurately characterize semantic variation. To address this, we investigate two approaches: one based on attention weights and another leveraging the more interpretable integrated gradients method, both designed to measure the influence of context tokens on the target token and to rank candidates by incorporating semantic similarity between the original and substituted sentences. Experiments on the LS07 and SWORDS datasets demonstrate that both approaches improve ranking performance.
查看此版本解读 →查看此版本原文 ↗