@inproceedings{k-etal-2025-goodmen,
    title = "goodmen @ {L}-{MT} Shared Task: A Comparative Study of Neural Models for {E}nglish-{H}indi Legal Machine Translation",
    author = "K, Deeraj S  and
      Suryanarayanan, Karthik  and
      Ingle, Yash  and
      Mishra, Pruthwik",
    editor = "Modi, Ashutosh  and
      Ghosh, Saptarshi  and
      Ekbal, Asif  and
      Goyal, Pawan  and
      Jain, Sarika  and
      Joshi, Abhinav  and
      Mishra, Shivani  and
      Datta, Debtanu  and
      Paul, Shounak  and
      Singh, Kshetrimayum Boynao  and
      Kumar, Sandeep",
    booktitle = "Proceedings of the 1st Workshop on NLP for Empowering Justice (JUST-NLP 2025)",
    month = dec,
    year = "2025",
    address = "Mumbai, India",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/ingest-ijcnlp-aacl/2025.justnlp-main.13/",
    pages = "127--132",
    ISBN = "979-8-89176-312-8",
    abstract = "In a massively multilingual country like India,providing legal judgments in understandablenative languages is essential for equitable jus-tice to all. The Legal Machine Translation(L-MT) shared task focuses on translating le-gal content from English to Hindi which is themost spoken language in India. We present acomprehensive evaluation of neural machinetranslation models for English-Hindi legal doc-ument translation, developed as part of the L-MT shared task. We investigate four multi-lingual and Indic focused translation systems.Our approach emphasizes domain specific fine-tuning on legal corpus while preserving statu-tory structure, legal citations, and jurisdic-tional terminology. We fine-tune two legalfocused translation models, InLegalTrans andIndicTrans2 on the English-Hindi legal paral-lel corpus provided by the organizers wherethe use of any external data is constrained.The fine-tuned InLegalTrans model achievesthe highest BLEU score of 0.48. Compara-tive analysis reveals that domain adaptationthrough fine-tuning on legal corpora signifi-cantly enhances translation quality for special-ized legal texts. Human evaluation confirmssuperior coherence and judicial tone preserva-tion in InLegalTrans outputs. Our best per-forming model is ranked 3rd on the test data."
}