ReflectMT: Internalizing Reflection for Efficient and High-Quality Machine Translation

• URL: http://arxiv.org/abs/2604.19144v1
• Date: Tue, 21 Apr 2026 06:48:41 GMT
• Title: ReflectMT: Internalizing Reflection for Efficient and High-Quality Machine Translation
• Abstract summary: We propose ReflectMT, a two-stage reflection internalization algorithm for machine translation.<n>Our approach develops the model's "translate-reflect-refine" capability through reinforcement learning.<n> Experimental results on datasets such as WMT24 demonstrate that our model's first-pass translations during inference outperform multi-step reasoning LRMs.
• Score: 63.89730738735446
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Recent years have witnessed growing interest in applying Large Reasoning Models (LRMs) to Machine Translation (MT). Existing approaches predominantly adopt a "think-first-then-translate" paradigm. Although explicit reasoning trajectories significantly enhance translation quality, they incur prohibitive inference costs and latency. To address these limitations, we propose ReflectMT, a two-stage reflection internalization algorithm for machine translation that employs a "translate-first-think-later" paradigm. Our approach develops the model's "translate-reflect-refine" capability through reinforcement learning. In the first stage, we cultivate the model's capacity for high-quality reflection and refinement, thereby enhancing its semantic comprehension and task-specific knowledge. In the second stage, we train the model to internalize the knowledge acquired during reflection. As a result, during inference, ReflectMT operates in a direct translation mode, producing high-quality translations on the first attempt without any explicit reasoning steps. Experimental results on datasets such as WMT24 demonstrate that our model's first-pass translations during inference outperform multi-step reasoning LRMs such as DeepSeek-R1 in both automatic metrics and GPT-based evaluation, achieving a 2.16-point improvement in GPT-based translation quality evaluation while reducing token consumption by 94.33%.

Related papers

• LatentMT: Machine Translation with Latent Reasoning [30.307790715793942]
  We introduce LatentMT, the first systematic study of latent-reasoning LoopLMs for machine translation.<n>Across 32 translation directions, LatentMT achieves performance comparable to models three to five times larger.<n>LatentMT requires lower training and inference compute than much larger non-latent-reasoning models.
  arXiv  Detail & Related papers  (2026-07-21T01:38:23Z)
• Incentivizing Parametric Knowledge via Reinforcement Learning with Verifiable Rewards for Cross-Cultural Entity Translation [68.85147984815778]
  We propose EA-RLVR (Entity-Anchored Reinforcement Learning with Verifiable Rewards) to incentivize the effective use of parametric knowledge.<n>EA-RLVR anchors supervision on a verifiable, entity-level reward signal and incorporates lightweight structural gates to stabilize optimization.<n>We evaluate EA-RLVR on XC-Translate and observe consistent improvements in both entity translation accuracy and out-of-domain generalization.
  arXiv  Detail & Related papers  (2026-04-18T07:15:43Z)
• Unlocking Reasoning Capability on Machine Translation in Large Language Models [57.60641851466707]
  Reasoning-oriented large language models (RLMs) achieve strong gains on tasks such as mathematics and coding by generating explicit intermediate reasoning.<n>We systematically evaluate several open- and closed-weights RLMs on the WMT24++ benchmark.<n>We find that enabling explicit reasoning consistently degrades translation quality across languages and models.
  arXiv  Detail & Related papers  (2026-02-16T14:05:59Z)
• Reflective Translation: Improving Low-Resource Machine Translation via Structured Self-Reflection [0.15229257192293197]
  Low-resource languages such as isiZulu and isiXhosa face persistent challenges in machine translation due to limited parallel data and linguistic resources.<n>Recent advances in large language models suggest that self-reflection, prompting a model to critique and revise its own outputs, can improve reasoning quality and factual consistency.<n>This paper introduces Reflective Translation, a prompt-based framework in which a model generates an initial translation, produces a structured self-critique, and then uses this reflection to generate a refined translation.
  arXiv  Detail & Related papers  (2026-01-27T18:37:09Z)
• TasTe: Teaching Large Language Models to Translate through Self-Reflection [82.83958470745381]
  Large language models (LLMs) have exhibited remarkable performance in various natural language processing tasks. We propose the TasTe framework, which stands for translating through self-reflection. The evaluation results in four language directions on the WMT22 benchmark reveal the effectiveness of our approach compared to existing methods.
  arXiv  Detail & Related papers  (2024-06-12T17:21:21Z)
• DUAL-REFLECT: Enhancing Large Language Models for Reflective Translation through Dual Learning Feedback Mechanisms [43.148203559785095]
  Large language models (LLMs) enhanced by self-reflection have achieved promising performance on machine translation. Existing self-reflection methods lack effective feedback information, limiting the translation performance. We introduce aREFLECT framework, leveraging the dual learning of translation tasks to provide effective feedback.
  arXiv  Detail & Related papers  (2024-06-11T13:10:39Z)
• Understanding and Addressing the Under-Translation Problem from the Perspective of Decoding Objective [72.83966378613238]
  Under-translation and over-translation remain two challenging problems in state-of-the-art Neural Machine Translation (NMT) systems. We conduct an in-depth analysis on the underlying cause of under-translation in NMT, providing an explanation from the perspective of decoding objective. We propose employing the confidence of predicting End Of Sentence (EOS) as a detector for under-translation, and strengthening the confidence-based penalty to penalize candidates with a high risk of under-translation.
  arXiv  Detail & Related papers  (2024-05-29T09:25:49Z)
• Thinker-DDM: Modeling Deliberation for Machine Translation with a Drift-Diffusion Process [14.600972421086219]
  We propose Thinker with the Drift-Diffusion Model to emulate human translators' dynamic decision-making under constrained resources.<n>We conduct experiments under the high-resource, low-resource, and commonsense translation settings using the WMT22 and CommonMT datasets.<n>We also perform additional analysis and evaluation on commonsense translation to illustrate the high effectiveness and efficacy of the proposed method.
  arXiv  Detail & Related papers  (2024-02-16T14:00:56Z)
• Towards Effective Disambiguation for Machine Translation with Large Language Models [65.80775710657672]
  We study the capabilities of large language models to translate "ambiguous sentences" Experiments show that our methods can match or outperform state-of-the-art systems such as DeepL and NLLB in four out of five language directions.
  arXiv  Detail & Related papers  (2023-09-20T22:22:52Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.