C.M. Downey

Dr. C.M. Downey is an Assistant Professor in Linguistics and Data Science at the University of Rochester, specializing in natural language processing for low-resource languages. His research develops methods for unsupervised learning, multilingual modeling, and transfer learning to address data scarcity in computational linguistics.

He received his PhD in Computational Linguistics from the University of Washington (2024), with dissertation work on segmental language models and cross-lingual transfer. Professional experience includes positions at Microsoft Research (cryptography and privacy) and Apple (AI/ML).

Research addresses fundamental challenges in language technology for under-resourced languages through novel neural architectures, including masked segmental models, hierarchical adaptation techniques, and communication-based translation frameworks. His contributions enable NLP applications for languages lacking large text corpora.

Analysis of publications shows consistent focus on multilingual embedding spaces (3 papers), segmentation/transduction models (2 papers), and low-resource adaptation techniques (2 papers), with recent emphasis on communication-inspired learning paradigms.

Current projects investigate dynamic vocabulary adaptation and emergent communication protocols for translation systems. Teaching includes graduate courses in computational linguistics and deep learning for NLP.