It Takes Two: A Dual Stage Approach for Terminology-Aware Translation

This paper introduces DuTerm, a novel two-stage architecture for terminology-constrained machine translation. Our system combines a terminology-aware NMT model, adapted via fine-tuning on large-scale synthetic data, with a prompt-based LLM for post-editing. The LLM stage refines NMT output and enforces terminology adherence. We evaluate DuTerm on English-to German, English-to-Spanish, and English-to-Russian with the WMT 2025 Terminology Shared Task corpus. We demonstrate that flexible, context-driven terminology handling by the LLM consistently yields higher quality translations than strict constraint enforcement. Our results highlight a critical trade-off, revealing that an LLM's work best for high-quality translation as context-driven mutators rather than generators.

• Publication year

  2025

• Venue

  Proceedings of the Tenth Conference on Machine Translation

• Publication date

  2025-11-07

• Fields of study

  Linguistics, Computer Science

• Identifiers
  DOI 10.18653/v1/2025.wmt-1.112arXiv 2511.07461
• External record

  Open on Semantic Scholar

• Source metadata

  Semantic Scholar

• No claims are published for this paper.

• No concepts are published for this paper.

• Findings of the WMT25 Terminology Translation Task: Terminology is Useful Especially for Good MTs

  2025cited by this paper
• Overview and challenges of machine translation for contextually appropriate translations

  2024cited by this paper
• Leveraging GPT-4 for Automatic Translation Post-Editing

  2023cited by this paper
• Understanding and Improving the Robustness of Terminology Constraints in Neural Machine Translation

  2023cited by this paper
• No Language Left Behind: Scaling Human-Centered Machine Translation

  2022cited by this paper
• CometKiwi: IST-Unbabel 2022 Submission for the Quality Estimation Shared Task

  2022cited by this paper
• Lexically Constrained Neural Machine Translation with Levenshtein Transformer

  2020cited by this paper
• Training Neural Machine Translation to Apply Terminology Constraints

  2019cited by this paper
• A Call for Clarity in Reporting BLEU Scores

  2018cited by this paper
• Google’s Multilingual Neural Machine Translation System: Enabling Zero-Shot Translation

  2016cited by this paper
• chrF: character n-gram F-score for automatic MT evaluation

  2015cited by this paper
• Neural Machine Translation by Jointly Learning to Align and Translate

  2014cited by this paper
• Bleu: a Method for Automatic Evaluation of Machine Translation

  2002cited by this paper

• Findings of the WMT25 Terminology Translation Task: Terminology is Useful Especially for Good MTs

  2025cites this paper
• An Empirical Analysis of Machine Translation for Expanding Multilingual Benchmarks

  2025cites this paper