TEaR: Improving LLM-based Machine Translation with Systematic Self-Refinement

中文速览

大语言模型虽然已经能做出高水平机器翻译，但译文仍常有遗漏、误译和表达不自然等问题，而且普通的自动润色缺少明确的错误判断和反馈。TEaR（Translate、Estimate、Refine）让同一个大模型先翻译，再按多维度翻译质量标准评估错误，最后依据具体反馈修改译文。实验覆盖10种语言、17个翻译方向和多种模型，结果显示TEaR在高低资源语言及不同语言组合上都稳定提升质量，整体优于多种已有后编辑方法，也更受人工评审偏好，其中错误评估反馈的准确性是效果好坏的关键。它把翻译、评价和修改串成了一个透明可解释的闭环，说明通用大模型不仅能翻译，还能借助自我检查有效改进译文，为低成本提升多语种机器翻译提供了实用思路。

摘要

Large Language Models (LLMs) have achieved impressive results in Machine Translation (MT). However, careful evaluations by human reveal that the translations produced by LLMs still contain multiple errors. Importantly, feeding back such error information into the LLMs can lead to self-refinement and result in improved translation performance. Motivated by these insights, we introduce a systematic LLM-based self-refinement translation framework, named TEaR, which stands for Translate, Estimate, and Refine, marking a significant step forward in this direction. Our findings demonstrate that 1) our self-refinement framework successfully assists LLMs in improving their translation quality across a wide range of languages, whether it’s from high-resource languages to low-resource ones or whether it’s English-centric or centered around other languages; 2) TEaR exhibits superior systematicity and interpretability; 3) different estimation strategies yield varied impacts, directly affecting the effectiveness of the final corrections. Additionally, traditional neural translation models and evaluation models operate separately, often focusing on singular tasks due to their limited capabilities, while general-purpose LLMs possess the capability to undertake both tasks simultaneously. We further conduct cross-model correction experiments to investigate the potential relationship between the translation and evaluation capabilities of general-purpose LLMs. Our code and data are available at https://github.com/fzp0424/self_correct_mt.

术语表

Large Language Models (LLMs)

大语言模型（LLM）

Machine Translation (MT)

机器翻译（MT）

TEaR

TEaR

Translate, Estimate, and Refine

翻译、评估与优化

self-refinement

自我优化

LLM-based self-refinement translation framework

基于LLM的自我优化翻译框架

translation quality

翻译质量

translation evaluation

翻译评估

error feedback

错误反馈

self-correction

自我纠错

cross-model correction

跨模型纠错

General Machine Translation Task

通用机器翻译任务

WMT 2023 Shared Task

WMT 2023共享任务

WMT23 Zh-En dataset

WMT23中英数据集

GPT-4

GPT-4

few-shot prompting

少样本提示

Multi-dimensional Quality Metrics (MQM)

多维质量指标（MQM）

COMET

COMET

Chain-of-Thought (CoT)

思维链（CoT）

translation post-editing

翻译译后编辑

Quality Estimation (QE)

质量估计（QE）

neural quality estimation models

神经质量估计模型

error span

错误片段

high-resource language

高资源语言

low-resource language

低资源语言

NLLB

NLLB

prompting strategy

提示策略

semantic-related metrics

语义相关指标