Thursday 23 Feb 2017: Can we machine-learn programming language semantics?
Dr. Dan Ghica - University of Birmingham
Harrison 170 14:30-15:30
Giving precise mathematical models of programming languages is a notoriously difficult problem. In this talk I will present and motivate the problem, with an emphasis on a particular semantic methodology called “game semantics”. I will show how game semantics has a more suitable structure for the use of machine learning techniques than other more common approaches such as domain-based or operational semantics. Finally I will show some preliminary experiments examining whether certain common features of game semantics are learnable using conventional machine learning algorithms.