PhD Student, Computer Science, BYU

Studying why language technology fails underrepresented communities, and how to fix it.

Isaac Thompson

I'm a PhD student in Computer Science at Brigham Young University, advised by Dr. Eric Ringger in the MATRIX Lab. My research focuses on low-resource machine translation, multilingual NLP, model interpretability, and cross-lingual evaluation.

Low-Resource MTInterpretabilityMultilingual NLPCross-Lingual Eval
ScholarGitHubLinkedInCV

Selected Publications

All research >
  1. 01

    AmanaMT: Translation Direction Predicts Automatic Metric Reliability

    Benchmarked 10 MT evaluation metrics across 730,748 segments and 33 languages. Translation direction is the strongest predictor of metric–human agreement (η² = 0.31), far ahead of morphological type (η² = 0.07). Also identifies a Simpson's Paradox in multi-year evaluation.

    PaperCode
    WMT 2026 @ EMNLP
  2. 02

    Bringing Mapudungun into the Modern MT Ecosystem

    Fine-tuned NLLB-200 on Mapudungun–Spanish across eight tokenization strategies. A novel Morfessor-VC segmentation matches a 5× larger model on arn→es (43.2 vs. 42.9 chrF++).

    PaperCode
    AmericasNLP 2026 @ ACL

Currently