mlsanityData to 14 Oct 2026

Faster Machine Translation Ensembling with Reinforcement Learning and Competitive Correction

arXiv 2501.15219 ↗First posted 25 Jan 2025cs.CLcited by 1

On the map: Deep reinforcement learning methods → RL training of language and agent models

Abstract (arXiv)

Ensembling neural machine translation (NMT) models to produce higher-quality translations than the L individual models has been extensively studied. Recent methods typically employ a candidate selection block (CSB) and an encoder-decoder fusion block (FB), requiring inference across textitall candidate models, leading to significant computational overhead, generally Omega(L). This paper introduces SmartGen, a reinforcement learning (RL)-based strategy that improves the CSB by selecting a small, fixed number of candidates and identifying optimal groups to pass to the fusion block for each input sentence. Furthermore, previously, the CSB and FB were trained independently, leading to suboptimal NMT performance. Our DQN-based SmartGen addresses this by using feedback from the FB block as a reward during training. We also resolve a key issue in earlier methods, where candidates were passed to the FB without modification, by introducing a Competitive Correction Block (CCB). Finally, we validate our approach with extensive experiments on English-Hindi translation tasks in both directions.

Key facts. Cited by 1 paper in the map (data through 14 Oct 2026). Placed in RL training of language and agent models, one of 15 directions in Deep reinforcement learning methods. Builds on 30 of its 45 references that are in the map. First posted on arXiv 25 Jan 2025.

Builds on

30 of 45 references.

References · 30

  1. 2015Deep Residual Learning for Image Recognitioncited by 52,096
  2. 2023GPT-4 Technical Reportcited by 18,396
  3. 2022Training language models to follow instructions with human feedbackcited by 15,971
  4. 2019Unsupervised Cross-lingual Representation Learning at Scalecited by 3,935
  5. 2013Playing Atari with Deep Reinforcement Learningcited by 3,274
  6. 2020Learning to summarize from human feedbackcited by 2,588
  7. 2020DeBERTa: Decoding-enhanced BERT with Disentangled Attentioncited by 2,181
  8. 2022BLOOM: A 176B-Parameter Open-Access Multilingual Language Modelcited by 1,984
  9. 2020mT5: A massively multilingual pre-trained text-to-text transformercited by 1,818
  10. 2018A Call for Clarity in Reporting BLEU Scorescited by 1,799
  11. 2022No Language Left Behind: Scaling Human-Centered Machine Translationcited by 1,312
  12. 2020Multilingual Denoising Pre-training for Neural Machine Translationcited by 1,035
18 more
  1. 2024Medusa: Simple LLM Inference Acceleration Framework with Multiple Decoding Headscited by 843
  2. 2023LLM-Blender: Ensembling Large Language Models with Pairwise Ranking and Generative Fusioncited by 564
  3. 2020Beyond English-Centric Multilingual Machine Translationcited by 561
  4. 2021The FLORES-101 Evaluation Benchmark for Low-Resource and Multilingual Machine Translationcited by 560
  5. 2024Aya Model: An Instruction Finetuned Open-Access Multilingual Language Modelcited by 295
  6. 2023Automatically Correcting Large Language Models: Surveying the landscape of diverse self-correction strategiescited by 238
  7. 2021Efficient Large Scale Language Modeling with Mixtures of Expertscited by 175
  8. 2023IndicTrans2: Towards High-Quality and Accessible Machine Translation Models for all 22 Scheduled Indian Languagescited by 152
  9. 2024LLM Inference Unveiled: Survey and Roofline Model Insightscited by 152
  10. 2021SimCLS: A Simple Framework for Contrastive Learning of Abstractive Summarizationcited by 132
  11. 2022SummaReranker: A Multi-Task Mixture-of-Experts Re-ranking Framework for Abstractive Summarizationcited by 61
  12. 2020Unbabel's Participation in the WMT20 Metrics Shared Taskcited by 56
  13. 2023Leveraging GPT-4 for Automatic Translation Post-Editingcited by 56
  14. 2024Airavata: Introducing Hindi Instruction-tuned LLMcited by 32
  15. 2023INSTRUCTSCORE: Explainable Text Generation Evaluation with Finegrained Feedbackcited by 22
  16. 2018Reinforcement Learning based Dynamic Model Selection for Short-Term Load Forecastingcited by 4
  17. 2019Adaptive Model Selection Framework: An Application to Airline Pricingcited by 3
  18. 2021Online and Scalable Model Selection with Multi-Armed Banditscited by 1

Built on by

Cited by year

Citing papers per year: 2025: 1.

1 paper.

  1. 2025In-Domain African Languages Translation Using LLMs and Multi-armed Banditscited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2017Reinforcement Learning for Bandit Neural Machine Translation with Simulated Human Feedback
  2. 2026Reference-Free Reinforcement Learning Fine-Tuning for MT: A Seq2Seq Perspective
  3. 2025SeqPO-SiMT: Sequential Policy Optimization for Simultaneous Machine Translation
  4. 2026PEGRL: Improving Machine Translation by Post-Editing Guided Reinforcement Learning
  5. 2025MT-R1-Zero: Advancing LLM-based Machine Translation via R1-Zero-like Reinforcement Learning
  6. 2024Fine-Grained Reward Optimization for Machine Translation using Error Severity Mappings
  7. 2025SSR-Zero: Simple Self-Rewarding Reinforcement Learning for Machine Translation
  8. 2024Reinforcement Learning for Edit-Based Non-Autoregressive Neural Machine Translation