I am a PhD candidate in Computer Science at ANU, supervised by Dirk Pattinson. My research asks what supervised learning looks like when errors are measured in p-adic and ultrametric spaces rather than ordinary Euclidean ones. I am particularly interested in when these structures produce simpler, more interpretable models, and what they change about expressivity and computational complexity.
I took a slightly indirect route to the PhD: high-school dropout, apprentice pipe-organ builder, Google site reliability engineer, startup CTO and, eventually, university lecturer. My earlier research ranges from an early programming language for simulating quantum computation to multilingual NLP across nearly 1,500 languages. I tend to be drawn to problems where several fields appear unrelated until the same mathematical structure turns up in all of them.
Key publications:
- Expressivity and NP-Hardness of p-Adic Linear Regression: https://doi.org/10.1134/S2070046626020081
- Linear Regression in p-Adic Metric Spaces: https://doi.org/10.1134/S2070046625040016
- Number Theory Meets Linguistics: https://aclanthology.org/2022.aacl-short.4/
- The Construction and Evaluation of the LEAFTOP Dataset: https://aclanthology.org/2022.lrec-1.536/