Generative Code Modeling with Graphs

M. Brockscmidt, M. Allamanis A. L. Gaunt, O. Polozov. ICLR 2019

[ArXiV] [OpenReview] [Code]      

Generative models forsource code are an interesting structured prediction problem, requiring to reason about both hard syntactic and semantic constraints as well as about natural, likely programs. We present a novel model for this problem that uses a graph to represent the intermediate state of the generated output. Our model generates code by interleaving grammar-driven expansion steps with graph augmentation and neural message passing steps. An experimental evaluation shows that our new model can generate semantically meaningful expressions, outperforming a range of strong baselines.