Lune

ACL Findings2026

Learning to Translate by Translating: Stabilizing the Dual Loop via Semantic-Aware Self-Evolution

Kui Liu, Mingming Yin, Zuoli Tang, Zihao Li, Chilin Fu, Xiaolu Zhang, Jun Zhou, Lixin Zou, Chenliang Li

Abstract

Despite the remarkable success of Large Language Models (LLMs) in Machine Translation (MT), the scarcity of high-quality parallel corpora and the prohibitive cost of their acquisition constrain scalability. To this end, we propose Learning to Translate by Translating (LTT), an LLM-driven dual-learning framework that enables autonomous translation, achieving an 80.42% performance improvement over the base model. By adapting the cycle-consistency principle to the generative paradigm, LTT eliminates the need for parallel data. It employs a robust semantic-aware reward function that balances adequacy with reconstruction fidelity, effectively mitigating the reward hacking issues inherent in traditional unsupervised MT. Relying solely on monolingual data, our 8B model consistently outperforms significantly larger models (70B+) in low-resource settings and achieves parity with state-of-the-art supervised baselines on mainstream benchmarks. LTT thus offers a scalable, data-efficient paradigm for autonomous machine translation.

Ask Lune

Know this paper in minutes, not an afternoon.

Lune can read every page of it and the top-tier papers that cite it, and gives you its claims, its evidence and its limits, each one cited to the paper it comes from. No skimming, no AI slop.

Questions to start from

Lune can draw on

the full paper · related work

Get the answer

Free to start. No credit card required.

Lineage

  • MT-R1-Zero: Advancing LLM-based Machine Translation via R1-Zero-like Reinforcement Learning cites A Paradigm Shift in Machine Translation: Boosting Translation Performance of Large Language Models
Builds onCited byLeft to right by year; the more cited, the higher.

Builds on8

  1. Self-Rewarding Language ModelsWeizhe Yuan, Richard Yuanzhe Pang, Kyunghyun Cho, Xian Li et al.ICML 2024 · 569 citations
  2. A Paradigm Shift in Machine Translation: Boosting Translation Performance of Large Language ModelsHaoran Xu, Young Jin Kim, Amr Sharaf, Hany Hassan AwadallaICLR 2024 · 122 citations
  3. In-context Examples Selection for Machine TranslationSweta Agrawal, Chunting Zhou, Mike Lewis, Luke Zettlemoyer et al.ACL Findings 2023 · 78 citations
  4. TOWER+: Bridging Generality and Translation Specialization in Multilingual LLMsRicardo Rei, Nuno Miguel Guerreiro, José Pombal, João Alves et al.ACL 2026 · 34 citations
  5. MT-R1-Zero: Advancing LLM-based Machine Translation via R1-Zero-like Reinforcement LearningZhaopeng Feng, Shaosheng Cao, Jiahan Ren, Jiayuan Su et al.EMNLP Findings 2025 · 9 citations

Related papers

  1. CYCLE-INSTRUCT: Fully Seed-Free Instruction Tuning via Dual Self-Training and Cycle ConsistencyZhanming Shen, Hao Chen, Yulei Tang, Shaolin Zhu et al.EMNLP 2025
  2. LLMs Are Zero-Shot Context-Aware Simultaneous TranslatorsRoman Koshkin, Katsuhito Sudoh, Satoshi NakamuraEMNLP 2024
  3. TRANS-ZERO: Self-Play Incentivizes Large Language Models for Multilingual Translation Without Parallel DataWei Zou, Sen Yang, Yu Bao, Shujian Huang et al.ACL Findings 2025 · 3 citations
  4. Self-Distillation for Model Stacking Unlocks Cross-Lingual NLU in 200+ LanguagesFabian David Schmidt, Philipp Borchert, Ivan Vulic, Goran GlavasEMNLP Findings 2024 · 2 citations
  5. Leveraging Unpaired Data for Vision-Language Generative Models via Cycle ConsistencyTianhong Li, Sangnie Bhardwaj, Yonglong Tian, Han Zhang et al.ICLR 2024 · 8 citations
Dusk over the sea between two cliffs drawn in fine vertical lines