Research

Research interests

My research focuses on machine translation for low-resource and endangered languages. In particular:

  • low-resource machine translation
  • endangered language documentation
  • natural language processing
  • data augmentation and synthetic data
  • pivot translation methods
  • interactive machine learning

Current work

I am a research assistant at the AI4 Communication Sciences Lab under the supervision of Dr. Emily Prud’hommeaux. My work is funded by an NSF REU grant focused on Formosan language corpora.

I investigate methods for improving translation among the Indigenous languages of Taiwan using data augmentation, pivot languages, and diverse corpora. This work uses mBART, NLLB, OpenNMT, Fairseq, and high-performance computing.

Publications

First-author publications

  1. 2026 Hunter Scheppat, Joshua K. Hartshorne, Sercan Koc, Éric Le Ferrand, and Emily Prud’hommeaux. “FormosanMT: A Multilingual Parallel Corpus of the Formosan Language Family.” Proceedings of the Fifteenth Language Resources and Evaluation Conference, pp. 3445–3455. doi · talk · conference page
  2. 2025 Hunter Scheppat, Joshua Hartshorne, Dylan Leddy, Éric Le Ferrand, and Emily Prud’hommeaux. “Integrating diverse corpora for training an endangered language machine translation system.” Proceedings of ComputEL-8, pp. 162–169. pdf · talk · demo

Co-authored publications

  1. 2026 “A Family Affair: Counterintuitive transfer effects in ASR for endangered languages.” Forthcoming in Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing (EMNLP 2026). Main Conference paper. Author list and paper coming soon.

Research experience

2026–present
Post-graduate research assistant
AI4 Communication Sciences Lab, MGH / BC / NSF
High-performance computing, machine translation, and ASR.
2023–present
Research assistant
AI4 Communication Sciences Lab, MGH / BC / NSF
NSF REU-funded Formosan MT research; synthetic data pipelines; evaluation of mBART, NLLB, and OpenNMT.
2023
Research assistant
Pushkin Project, Boston College
Behavioral experiments, JavaScript and Selenium testing, and OpenSSF compliance.

Selected presentations

  • LREC 2026. “FormosanMT: A Multilingual Parallel Corpus of the Formosan Language Family.” Oral presentation.
  • ComputEL-8. “Integrating diverse corpora for training an endangered language machine translation system.” Oral presentation.