Research Interests

To understand and reason about code, we (=humans) combine its highly structured, formal nature with ambiguous information in the identifier names, comments, and coding idioms. Can we devise computational methods that replicate this form of reasoning? I research machine learning models and methods that "understand" and generate code. My objective is to invent better machine learning methods for semi-structured reasoning, and to inspire novel software engineering tools that will assist developers in their work.

I am currently a principal researcher at Microsoft Research in Cambridge, UK and part of the Deep Program Understanding project in the Machine Intelligence group.

Highlighted Publications

• Typilus: Neural Type Hints
  M. Allamanis, E. T. Barr, S. Ducousso, Z. Gao. PLDI 2020
• Fast and Memory-Efficient Neural Code Completion
  A. Svyatkovskiy, S. Lee, A. Hadjitofi, M. Riechert, J. Franco, M. Allamanis. 2020
• Learning to Represent Edits
  P. Yin, G. Neubig, M. Allamanis, M. Brockschmidt, A. L. Gaunt. ICLR 2019
• The Adverse Effects of Code Duplication in Machine Learning Models of Code
  M. Allamanis. SPLASH Onward! 2019
• A Survey of Machine Learning for Big Code and Naturalness
  M. Allamanis, E. T. Barr, P. Devanbu, C. Sutton. ACM Computing Surveys 2018
• Learning to Represent Programs with Graphs
  M. Allamanis, M. Brockscmidt, M. Khademi. ICLR 2018

Full list of publications