--- title: "Salute the Classic — paper by Pang et al. (2025)" description: "The paper asks whether the six classic obstacles for neural MT identified by Koehn and… 62.0 BLEU. 2 references, cited by 32 papers here." authors: ["Jianhui Pang", "Fanghua Ye", "Derek Fai Wong", "Dian Yu", "Shuming Shi", "Zhaopeng Tu", "Longyue Wang"] year: 2025 venue: "Transactions of the Association for Computational Linguistics" source: "https://aclanthology.org/2025.tacl-1.4/" doi: "10.1162/tacl_a_00730" licence: "CC BY 4.0 (ACL Anthology)" licence_url: "https://creativecommons.org/licenses/by/4.0/" summary_licence: "Elamur Cosmic License" canonical: "https://research.elamur.ai/free/en/papers/pang-2025-salute-classic-revisiting-challenges-machine/" --- [Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/pang-2025-salute-classic-revisiting-challenges-machine/) [home](https://md.research.elamur.ai/free/en/) / [papers](https://md.research.elamur.ai/free/en/papers/) / [2025](https://md.research.elamur.ai/free/en/papers/2025/page/2/) / Salute the Classic: Revisiting Challenges of Machine Translation in the Age of Large Language Models # Salute the Classic: Revisiting Challenges of Machine Translation in the Age of Large Language Models **Authors:** [Jianhui Pang](https://md.research.elamur.ai/free/en/authors/jianhui-pang/), [Fanghua Ye](https://md.research.elamur.ai/free/en/authors/fanghua-ye/), [Derek Fai Wong](https://md.research.elamur.ai/free/en/authors/derek-f-wong/), [Dian Yu](https://md.research.elamur.ai/free/en/authors/dian-yu/), [Shuming Shi](https://md.research.elamur.ai/free/en/authors/shuming-shi/), [Zhaopeng Tu](https://md.research.elamur.ai/free/en/authors/zhaopeng-tu/) and [Longyue Wang](https://md.research.elamur.ai/free/en/authors/longyue-wang/) · **Year:** 2025 · **Venue:** [Transactions of the Association for Computational Linguistics](https://md.research.elamur.ai/free/en/venues/tacl/) ([Transactions of the Association for Computational Linguistics (TACL) 2025](https://md.research.elamur.ai/free/en/venues/tacl/2025/)) · **Publisher:** [The MIT Press](https://md.research.elamur.ai/free/en/publishers/the-mit-press/) · **Licence:** [CC BY 4.0 (ACL Anthology)](https://creativecommons.org/licenses/by/4.0/) **Links:** [Publisher's page](https://aclanthology.org/2025.tacl-1.4/) · [PDF](https://aclanthology.org/2025.tacl-1.4.pdf) · [DOI 10.1162/tacl\_a\_00730](https://doi.org/10.1162/tacl_a_00730) · [OpenAlex](https://openalex.org/W4406237009) **Citations:** 33 citations (OpenAlex, 2026-10-14) · [32](https://md.research.elamur.ai/free/en/papers/pang-2025-salute-classic-revisiting-challenges-machine/cited-by/) in our library, of the 149,439 papers whose bibliographies we read ## Summary The paper asks whether the six classic obstacles for neural MT identified by Koehn and Knowles (2017) still apply when translation is done by decoder-only LLMs. The six are domain mismatch, parallel data volume, rare words, long sentences, attention-based word alignment, and beam search. The authors fine-tune Llama2-7B and Llama2-13B for German-to-English in two ways. LLM-SFT uses bilingual pairs mixed with Alpaca. LLM-CPT-SFT adds continued pretraining on concatenated pairs before Alpaca tuning. Both are compared with Fairseq Transformer (Enc2Dec) baselines on WMT23 and OPUS multi-domain data, scored with BLEU and COMET-DA. LLMs largely remove the need for large parallel corpora in well-resourced languages and handle long inputs well. Domain shift and rare words remain unsolved, and alignment and beam search matter less in the LLM setting. The study also identifies three LLM-era problems: - slow inference, - weak translation into languages underrepresented in pretraining, such as Ukrainian and Hebrew, - automatic metrics that disagree with human judgments. The work provides an updated research agenda for LLM-based MT, which is relevant to low-resource directions such as English-to-Ukrainian. ## Key points - **Setup:** Training draws up to 20M pairs from roughly 300M WMT23 De-En pairs. Testing uses generaltest2023 to avoid data leakage. Domain experiments use OPUS multi-domain sets. - **Domain results:** In the law domain, LLM-SFT reaches 62.0 BLEU and 88.0 COMET-DA, about 3.0 BLEU and 4.9 COMET-DA above Enc2Dec. Moving from Koran to Law, however, costs up to 40.0 BLEU and 14.4 COMET-DA. - **Long inputs and speed:** LLMs translate sentences of about 80 words well and manage documents of up to 512 words. Their inference was about 100 times slower than Enc2Dec in the authors' experiments. - **Ukrainian:** Ukrainian makes up about 0.07% of Llama2 pretraining data, and En→Uk quality jumps between 100k and 500k parallel pairs. ALMA-7B and TowerInstruct-7B reach 88.3 and 87.0 COMET-DA on En→Uk without Ukrainian pretraining, which the authors attribute to Russian data. - **Evaluation gap:** On the WMT23 literary Zh-En test, human scores (MQM) and automatic scores (d-BLEU) for GPT-4, a document-tuned Llama2-7B, and Google Translate correlate at Pearson −0.53. *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 - [Tencent](https://md.research.elamur.ai/free/en/companies/tencent/) — Company, Shenzhen, China · 163 works · 21,032 citations - [University College London](https://md.research.elamur.ai/free/en/universities/university-college-london/) — University, London, United Kingdom · 107 works · 22,923 citations - [University of Macau](https://md.research.elamur.ai/free/en/universities/university-of-macau/) — University, Macao, China · 43 works · 2,368 citations 📚 Shelves (4): [🖥️ Computer Science](https://md.research.elamur.ai/free/en/shelves/computer-science/) · 18,614 works · [🤖 Artificial Intelligence](https://md.research.elamur.ai/free/en/shelves/computer-science/artificial-intelligence/) · 17,965 works · [🎭 Arts and Humanities](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/) · 17,450 works · [🗣️ Language and Linguistics](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/language-and-linguistics/) · 17,296 works 🧭 Topics (1): [🔁 Machine translation](https://md.research.elamur.ai/free/en/topics/machine-translation/) · 1,105 works ## Cite Cite as: Jianhui Pang, Fanghua Ye, Derek Fai Wong, Dian Yu, Shuming Shi, Zhaopeng Tu and Longyue Wang. 2025. Salute the Classic: Revisiting Challenges of Machine Translation in the Age of Large Language Models. Transactions of the Association for Computational Linguistics. The paper itself, not this page. BibTeX from the library's record of the work. ```bibtex @article{pang2025salute, title = {{Salute the Classic: Revisiting Challenges of Machine Translation in the Age of Large Language Models}}, author = {Jianhui Pang, Fanghua Ye, Derek Fai Wong, Dian Yu, Shuming Shi, Zhaopeng Tu and Longyue Wang}, journal = {Transactions of the Association for Computational Linguistics}, year = {2025}, doi = {10.1162/tacl\_a\_00730}, url = {https://doi.org/10.1162/tacl_a_00730} } ``` - [← previous Nemotron-CC: Transforming Common Crawl into a Refined Long-Horizon Pretraining Dataset](https://md.research.elamur.ai/free/en/papers/su-2025-nemotron-cc-transforming-common-crawl/) - [next → A Comprehensive Survey on Long Context Language Modeling](https://md.research.elamur.ai/free/en/papers/liu-2025-comprehensive-survey-long-context-language/) - [↑ back to papers of 2025](https://md.research.elamur.ai/free/en/papers/2025/page/2/) ## References 2 references, 2 of them with a card of ours. ↗ leads to the work. 1. Amey Agrawal, Ashish Panwar, Jayashree Mohan, Nipun Kwatra, Bhargav S. Gulavani, and Ramachandran Ramjee. 2023. [Sarathi: Efficient llm inference by piggybacking decodes with chunked prefills](https://md.research.elamur.ai/free/en/papers/agrawal-2023-sarathi-efficient-llm-inference-piggybacking/). arXiv preprint arXiv:2308.16369. 2. The above findings highlight the importance of advancing evaluation methodologies that align with human preferences for LLM translation. \#\# Conclusions Roee Aharoni and Yoav Goldberg. 2020. [Unsupervised domain clusters in pretrained language models](https://md.research.elamur.ai/free/en/papers/aharoni-2020-unsupervised-domain-clusters-pretrained-language/). In Proceedings of the 58th Annual Meeting of the A… *Updated 2 November 2026*