Elamur Research⌕
☰
home / papers / 2026 / AdaDPI: Document-level Translation Adaptive Agent via Dynamic Parametric Internalization

AdaDPI: Document-level Translation Adaptive Agent via Dynamic Parametric Internalization

Paper · 2026 · Hong Ren, Liting Deng, Shaolin Zhu et al.

Summary

Document-level machine translation (DocMT) with LLMs often loses track of terminology, pronoun referents and style across long texts. Existing remedies either enlarge the context window or, like the agents DELTA and GRAFT, attach retrieved memory to the prompt. Both treat earlier context as fixed external input, which dilutes attention and slows inference. AdaDPI instead writes document knowledge into a small LoRA-style adapter during inference, while the base LLM stays frozen. A linguistic uncertainty monitor (LUM) uses token-level entropy and how spread out attention is to detect when the model lacks the context it needs. Only then does a context-to-parameter integrator (CPI) turn retrieved constraints into a few gradient updates. On the GuoFeng webnovel and IWSLT2017 benchmarks, the system improves consistency by over 5 points versus state-of-the-art baselines. It nearly matches the quality of adapting at every sentence while adding only modest latency. As the document progresses, the agent depends less and less on retrieval, which suggests a practical route to more consistent long-form translation.

Key points

  • Consistency: On GuoFeng En→Zh, Terminology Recall reaches 98.2%, and Pronoun Accuracy is 3.3% higher than DELTA.
  • Efficiency (single A100):
    • Static inference: 142 ms per sentence, 79.85 COMET.
    • Adapting at every sentence: 638 ms, 85.50 COMET.
    • AdaDPI: 175 ms (+23%), 85.42 COMET.
    • LUM triggers updates for only 18.4% of sentences.
  • Reduced reliance on retrieval: The LUM trigger rate falls from about 35% to under 10% over a document, while COMET keeps rising.
  • Ablations:
    • A randomly initialized adapter causes performance to collapse.
    • Updating only attention weights performs about as well as full updating; updating only the MLP layers costs 0.65 d-BLEU.
    • Five gradient steps per constraint reduce COMET by 1.62, a sign of catastrophic forgetting.
  • Generality: Gains hold on both Qwen and LLaMA-3-8B backbones, outperforming zero-shot, 5-shot ICL and LoRA SFT baselines.

Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.

 Institutions

🎓Tianjin UniversityTianjin · China41 works · 1,128 citations

 Fields, topics and concepts

📚 Shelves 4The field of science this work belongs to (OpenAlex fields).Computer Science · 16,667 worksArtificial Intelligence · 16,194 worksArts and Humanities · 16,039 worksLanguage and Linguistics · 15,891 works

🧭 Topics 1Research subjects and languages the work is about.Machine translation · 1,087 works

Cite

Cite as Hong Ren, Liting Deng, Shaolin Zhu and Deyi Xiong. 2026. AdaDPI: Document-level Translation Adaptive Agent via Dynamic Parametric Internalization. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers).

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.18653/v1/2026.acl-long.1758

@inproceedings{ren2026adadpi,
  title = {{AdaDPI: Document-level Translation Adaptive Agent via Dynamic Parametric Internalization}},
  author = {Hong Ren, Liting Deng, Shaolin Zhu and Deyi Xiong},
  booktitle = {Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)},
  year = {2026},
  doi = {10.18653/v1/2026.acl-long.1758},
  url = {https://doi.org/10.18653/v1/2026.acl-long.1758}
}

References

36 references, 10 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Duarte M Alves, José Pombal, Nuno M Guerreiro, Pedro H Martins, João Alves, Amin Farajian, Ben Peters, Ricardo Rei, Patrick Fernandes, Sweta Agrawal, Pierre Colombo, José G.C. de Souza, and André F.T. Martins. 2024. Tower: An open multilingual large language model for translation-related tasks. arXi…
  2. Guangsheng Bao, Yue Zhang, Zhiyang Teng, Boxing Chen, and Weihua Luo. 2021. G-transformer for document-level machine translation. In Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Vol…
  3. Mauro Cettolo, Marcello Federico, Luisa Bentivogli, Jan Niehues, Sebastian Stüker, Katsuhito Sudoh, Koichiro Yoshino, and Christian Federmann. 2017. Overview of the iwslt 2017 evaluation campaign. In Proceedings of the 14th International Conference on Spoken Language Translation, pages 2–14.
  4. Longze Chen, Ziqiang Liu, Wanwei He, Yinhe Zheng, Hao Sun, Yunshui Li, Run Luo, and Min Yang. 2024. Long context is not long at all: A prospector of long-dependency data for large language models. In Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: L… find
  5. Pranjal Chitale, Jay Gala, and Raj Dabre. 2024. An empirical study of in-context learning in LLMs for machine translation. In Findings of the Association for Computational Linguistics: ACL 2024, pages 7384–7406. find
  6. Marta R Costa-Jussà, James Cross, Onur Çelebi, Maha Elbayad, Kenneth Heafield, Kevin Heffernan, Elahe Kalbassi, Janice Lam, Daniel Licht, Jean Maillard, and 1 others. 2022. No language left behind: Scaling human-centered machine translation. arXiv preprint arXiv:2207.04672. find
  7. Menglong Cui, Jiangcun Du, Shaolin Zhu, and Deyi Xiong. 2024. Efficiently exploring large language models for document-level machine translation with in-context learning. In Findings of the Association for Computational Linguistics ACL 2024, pages 10885–10897. find
  8. Tianyu Dong, Bo Li, Jinsong Liu, Shaolin Zhu, and Deyi Xiong. 2025. Mlas-lora: language-aware parameters detection and lora-based knowledge transfer for multilingual machine translation. In Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers… find
  9. Xiaoyi Dong, and 1 others. 2024. Mmlongbench-doc: Benchmarking long-context document understanding with visualizations. Advances in Neural Information Processing Systems, 37:95963–96010. find
  10. Himanshu Dutta, Sunny Manchanda, Prakhar Bapat, Meva Ram Gurjar, and Pushpak Bhattacharyya. 2025. Graft: A graph-based flow-aware agentic framework for document-level machine translation. In Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing: Industry Track, pages… find
  11. Sameen Maruf, Fahimeh Saleh, and Gholamreza Haffari. 2021. A survey on document-level neural machine translation: Methods and evaluation. ACM Computing Surveys (CSUR), 54(2):1–36. find
  12. Daya Guo, Dejian Yang, Haowei Zhang, Junxiao Song, Ruoyu Zhang, Runxin Xu, Qihao Zhu, Shirong Ma, Peiyi Wang, Xiao Bi, and 1 others. 2025. Deepseek-r1: Incentivizing reasoning capability in LLMs via reinforcement learning. arXiv preprint arXiv:2501.12948.
  13. Edward J Hu, Yelong Shen, Phillip Wallis, Zeyuan Allen-Zhu, Yuanzhi Li, Shean Wang, Lu Wang, Weizhu Chen, and 1 others. 2022. Lora: Low-rank adaptation of large language models. ICLR, 1(2):3.
  14. Baijun Ji, Xiangyu Duan, Zhenyu Qiu, Tong Zhang, Junhui Li, Hao Yang, and Min Zhang. 2024. Submodular-based in-context example selection for LLMs-based machine translation. In Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC… find
  15. Marzena Karpinska and Mohit Iyyer. 2023. Large language models effectively leverage document-level context for literary translation, but critical errors persist. In Proceedings of the Eighth Conference on Machine Translation, pages 419–451.
  16. Yachao Li, Junhui Li, Jing Jiang, and Min Zhang. 2024. Enhancing document-level translation of large language model via translation mixed-instructions. arXiv preprint arXiv:2401.08088. — [arXiv:2401.08088](https://arxiv.org/abs/2401.08088) find
  17. Nelson F Liu, Kevin Lin, John Hewitt, Ashwin Paranjape, Michele Bevilacqua, Fabio Petroni, and Percy Liang. 2024. Lost in the middle: How language models use long contexts. Transactions of the Association for Computational Linguistics, 12:157–173.
  18. Yinhan Liu, Jiatao Gu, Naman Goyal, Xian Li, Sergey Edunov, Marjan Ghazvininejad, Mike Lewis, and Luke Zettlemoyer. 2020. Multilingual denoising pre-training for neural machine translation. Transactions of the Association for Computational Linguistics, 8:726–742. find
  19. Xinglin Lyu, Junhui Li, Zhengxian Gong, and Min Zhang. 2021. Encouraging lexical translation consistency for document-level neural machine translation. In Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing, pages 3265–3277. find
  20. Yubo Ma, Yuhang Zang, Liangyu Chen, Meiqi Chen, Yizhu Jiao, Xinze Li, Xinyuan Lu, Ziyu Liu, Yan Ma, Proyag Pal, Alexandra Birch, and Kenneth Heafield. 2024. Document-level machine translation with large-scale public parallel corpora. In Proceedings of the 62nd Annual Meeting of the Association for C… find
  21. Miguel Moura Ramos, Patrick Fernandes, Sweta Agrawal, and André FT Martins. 2025. Multilingual contextualization of large language models for document-level machine translation. arXiv preprint arXiv:2504.12140. — [arXiv:2504.12140](https://arxiv.org/abs/2504.12140) find
  22. Lei Tang, Jinghui Qin, Wenxuan Ye, Hao Tan, and Zhijing Yang. 2025. Adaptive few-shot prompting for machine translation with pre-trained language models. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 39, pages 25255–25263. find
  23. Gideon Toury. 2012. Descriptive translation studies: And beyond. find
  24. Jiaan Wang, Fandong Meng, Yingxue Zhang, and Jie Zhou. 2024a. Retrieval-augmented machine translation with unstructured knowledge. arXiv preprint arXiv:2412.04342. find
  25. Longyue Wang, Siyou Liu, Minghao Wu, Wenxiang Jiao, Xing Wang, Jiahao Xu, Zhaopeng Tu, Liting Zhou, Yan Gu, Weiyu Chen, Philipp Koehn, Andy Way, and Yulin Yuan. 2024b. Findings of the wmt 2024 shared task on discourse-level literary translation. In Proceedings of the Ninth Conference on Machine Tran… find
  26. Longyue Wang, Chenyang Lyu, Tianbo Ji, Zhirui Zhang, Dian Yu, Shuming Shi, and Zhaopeng Tu. 2023a. Document-level machine translation with large language models. In Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing, pages 16646–16661.
  27. Longyue Wang, Zhaopeng Tu, Yan Gu, Siyou Liu, Dian Yu, Qingsong Ma, Chenyang Lyu, Liting Zhou, Chao-Hong Liu, Yufeng Ma, and 1 others. 2023b. Findings of the wmt 2023 shared task on discourse-level literary translation: A fresh orb in the cosmos of LLMs. In Proceedings of the Eighth Conference on Ma… find
  28. Yutong Wang, Jiali Zeng, Xuebo Liu, Derek F Wong, Fandong Meng, Jie Zhou, and Min Zhang. 2024c. Delta: An online document-level translation agent based on multi-level memory. arXiv preprint arXiv:2410.08143. — [arXiv:2410.08143](https://arxiv.org/abs/2410.08143) find
  29. Minghao Wu, Thuy-Trang Vu, Lizhen Qu, George Foster, and Gholamreza Haffari. 2024. Adapting large language models for document-level machine translation. arXiv preprint arXiv:2401.06468. — [arXiv:2401.06468](https://arxiv.org/abs/2401.06468) find
  30. Qiying Yu, Zheng Zhang, Ruofei Zhu, Yufeng Yuan, Xiaochen Zuo, Yu Yue, Weinan Dai, Tiantian Fan, Gaohong Liu, Lingjun Liu, and 1 others. 2025. Dapo: An open-source LLM reinforcement learning system at scale. arXiv preprint arXiv:2503.14476.
  31. Guibin Zhang, Muxin Fu, and Shuicheng Yan. 2025a. Memgen: Weaving generative latent memory for self-evolving agents. arXiv preprint arXiv:2509.24704. — [arXiv:2509.24704](https://arxiv.org/abs/2509.24704) find
  32. Wen Zhang, Long Jin, Yushan Zhu, Jiaoyan Chen, Zhiwei Huang, Junjie Wang, Yin Hua, Lei Liang, and Huajun Chen. 2025b. Trustuqa: A trustful framework for unified structured data question answering. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 39, pages 25931–25939. find
  33. Zijian Zhou, Ao Qu, Zhaoxuan Wu, Sunghwan Kim, Alok Prakash, Daniela Rus, Jinhua Zhao, Bryan Kian Hsiang Low, and Paul Pu Liang. 2025. Mem1: Learning to synergize memory and reasoning for efficient long-horizon agents. arXiv preprint arXiv:2506.15841. — [arXiv:2506.15841](https://arxiv.org/abs/2506.15841) find
  34. Shaolin Zhu, Menglong Cui, and Deyi Xiong. 2024a. Towards robust in-context learning for machine translation with large language models. In Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024), pages 16619–16629. find
  35. Shaolin Zhu, Leiyu Pan, Bo Li, and Deyi Xiong. 2024b. Landermt: Dectecting and routing language-aware neurons for selectively finetuning LLMs to machine translation. In Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 12135–12148. find
  36. Wenhao Zhu, Hongyi Liu, Qingxiu Dong, Jingjing Xu, Shujian Huang, Lingpeng Kong, Jiajun Chen, and Lei Li. 2024c. Multilingual machine translation with large language models: Empirical results and analysis. In Findings of the association for computational linguistics: NAACL 2024, pages 2765–2781.

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/ren-2026-adadpi-document-level-translation-adaptive/