Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Cross-Preference Learning for Sentence-Level and Context-Aware Machine Translation

Ying Li, Xinglin Lyu, Junhui Li, Jinlong Yang, Hengchao Shang, Min Zhang, Shimin Tao, Daimeng Wei

Abstract

Context-aware machine translation (MT) leverages document-level information, yet it does not consistently outperform sentence-level MT, as contextual signals are unevenly beneficial across sentences. Existing training objectives do not explicitly model this variability, limiting a model's ability to adaptively exploit context. In this paper, we propose Cross-Preference Learning (CPL), a preference-based training framework that explicitly captures the complementary benefits of sentence-level and context-aware MT. CPL achieves this by integrating both intra- and cross-condition preferences into the preference optimization objective. The introduction of intra- and cross-condition preferences provides explicit supervision on when and how contextual information improves translation quality. We validate the proposed approach on several public context-aware MT tasks using multiple models, including Qwen3-4B, Qwen3-8B, and Llama-3-8B. Experimental results demonstrate consistent improvements in translation quality and robustness across both input conditions, achieved without any architectural modifications.

Cross-Preference Learning for Sentence-Level and Context-Aware Machine Translation

Abstract

Context-aware machine translation (MT) leverages document-level information, yet it does not consistently outperform sentence-level MT, as contextual signals are unevenly beneficial across sentences. Existing training objectives do not explicitly model this variability, limiting a model's ability to adaptively exploit context. In this paper, we propose Cross-Preference Learning (CPL), a preference-based training framework that explicitly captures the complementary benefits of sentence-level and context-aware MT. CPL achieves this by integrating both intra- and cross-condition preferences into the preference optimization objective. The introduction of intra- and cross-condition preferences provides explicit supervision on when and how contextual information improves translation quality. We validate the proposed approach on several public context-aware MT tasks using multiple models, including Qwen3-4B, Qwen3-8B, and Llama-3-8B. Experimental results demonstrate consistent improvements in translation quality and robustness across both input conditions, achieved without any architectural modifications.

Paper Structure

This paper contains 29 sections, 5 equations, 9 figures, 12 tables.

Table of Contents

  1. Introduction
  2. Methodology
  3. Sentence- and Context-Conditioned Translation
  4. Cross-Preference Learning
  5. Preference-Based Optimization Background
  6. Intra-Condition Preference Learning
  7. Cross-Condition Preference Learning
  8. CPL objective
  9. Experimentation
  10. Experimental Settings
  11. Main Results
  12. Discussion
  13. Does CPL Learn When to Use Context?
  14. Analysis with Additional Quality Metrics
  15. Evaluation with GPT-as-Judge
  16. ...and 14 more sections

Figures (9)

  • Figure 1: Distribution of COMET differences between sentence-level (Sent) and context-aware (Ctx) English–German translations produced by Qwen3-8B.
  • Figure 2: Illustration of our approach.
  • Figure 3: Distribution (%) of sentences by COMET differences between context-aware and sentence-level translations. CB/B/P/W/CW = Clearly Better, Better, On Par, Worse, Clearly Worse.
  • Figure 4: Case studies illustrating that CPL benefits from informative context.
  • Figure 5: Prompt used for context-dependency evaluation of sentence-level translation.
  • ...and 4 more figures