跳到正文
arXiv cs.CL· Terry Yue Zhuo, Qiongkai Xu, Xuanli He, Trevor Cohn·· 2 天前AI 评分26

重新审视往返翻译在机器翻译评估中的作用

Rethinking Round-Trip Translation for Machine Translation Evaluation

AI 导读

研究发现往返翻译可用于无参考的机器翻译自动评估。对 SMT 评估的重新审视表明,其长期被误解的根源在于复制机制,移除该机制后往返翻译分数能恰当反映正向翻译性能。往返翻译可用于预测正向翻译分数、提升质量估计模型表现,并通过跨系统验证识别共享任务中的对抗性竞争者。

正文 · 原文

View PDF HTML (experimental)

Abstract:Automatic evaluation on low-resource language translation suffers from a deficiency of parallel corpora. Round-trip translation could be served as a clever and straightforward technique to alleviate the requirement of the parallel evaluation corpus. However, there was an observation of obscure correlations between the evaluation scores by forward and round-trip translations in the era of statistical machine translation (SMT). In this paper, we report the surprising finding that round-trip translation can be used for automatic evaluation without the references. Firstly, our revisit on the round-trip translation in SMT evaluation unveils that its long-standing misunderstanding is essentially caused by copying mechanism. After removing copying mechanism in SMT, round-trip translation scores can appropriately reflect the forward translation performance. Then, we demonstrate the rectification is overdue as round-trip translation could benefit multiple machine translation evaluation tasks. To be more specific, round-trip translation could be used i) to predict corresponding forward translation scores; ii) to improve the performance of the recently advanced quality estimation model; and iii) to identify adversarial competitors in shared tasks via cross-system verification.
Comments: Accepted to Findings of ACL 2023
Subjects: Computation and Language (cs.CL)
Cite as: arXiv:2209.07351 [cs.CL]
  (or arXiv:2209.07351v4 [cs.CL] for this version)
  https://doi.org/10.48550/arXiv.2209.07351

arXiv-issued DOI via DataCite

Submission history

From: Terry Yue Zhuo [view email]
[v1] Thu, 15 Sep 2022 15:06:20 UTC (273 KB)
[v2] Tue, 13 Dec 2022 09:51:19 UTC (1,069 KB)
[v3] Mon, 15 May 2023 11:33:20 UTC (1,070 KB)
[v4] Fri, 13 Nov 2026 11:29:10 UTC (919 KB)

来源:arXiv cs.CL · arxiv.org