arXiv cs.CL· Javier Garcia Gilabert, Rico Sennrich, Andr\'e F. T. Martins·· 12 小时前AI 评分27
更多搜索何时有用？机器翻译质量感知解码中搜索效用的实证研究
When is more search useful? An Empirical Study of Search Utility in Quality-Aware Decoding for Machine Translation
AI 导读
针对质量感知解码（QAD）对每个源句固定分配相同候选预算 N 的假设，实证研究发现句子间从额外候选中获得的质量增益差异显著。研究用总缩放增益、边际缩放增益和采样饱和点三个量刻画搜索效用，发现句级信号与总缩放增益相关但对饱和点信息量有限，而当前候选池的统计特性可有效预测边际增益。这些信号能识别出选择性增加搜索时收益更大的句子。
正文 · 原文
Abstract:Quality-Aware Decoding (QAD) improves machine translation through test-time scaling, generating multiple candidate translations and selecting the final output with a quality-aware decoder. Usually the same candidate budget N is spent on every source sentence, assuming that all inputs require the same amount of search. We show that this assumption does not hold and that the quality gain obtained from additional candidates varies across sentences. We study search utility through three quantities: total scaling gain, how much quality a sentence can gain from a much larger candidate pool; marginal scaling gain, the expected gain from generating one additional batch of candidates; and the sampling saturation point, the candidate budget required to realize most of the total scaling gain. We investigate two sources of information for predicting these quantities: signals available before candidate-set expansion, and statistics computed from the candidate pool accumulated during search. We find that sentence-level signals correlate with total scaling gain but are less informative about the sampling saturation point. In contrast, properties of the current candidate pool provide useful signals of marginal scaling gain. Finally, we show that these signals can identify sentences that obtain larger gains when additional search is selectively applied.
| Subjects: | Computation and Language (cs.CL) |
| Cite as: | arXiv:2610.32679 [cs.CL] |
| (or arXiv:2610.32679v1 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2610.32679 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Javier García Gilabert [view email]
[v1]
Fri, 23 Oct 2026 14:15:48 UTC (3,589 KB)
来源：arXiv cs.CL · arxiv.org