Deep Learning On Code with an Unbounded Vocabulary

EasyChair Preprint 466

Abstract

A major challenge when using techniques from Natural Language Processing for supervised learning on computer program source code is that many words in code are neologisms. Reasoning over such an unbounded vocabulary is not something NLP methods are typically suited for. We introduce a deep model that contends with an unbounded vocabulary (at training or test time) by embedding new words as nodes in a graph as they are encountered and processing the graph with a Graph Neural Network.

Keyphrases: Graph Neural Network, Graph Neural Networks, Learning Representation, Natural Language Processing, abstract syntax tree, ast augmented ast, augmented ast, control flow, deep learning, fixed vocabulary, machine learning, neural network, source code, unbounded vocabulary, variable naming task

@booklet{EasyChair:466,
  author    = {Milan Cvitkovic and Badal Singh and Anima Anandkumar},
  title     = {Deep Learning On Code with an Unbounded Vocabulary},
  doi       = {10.29007/bc6w},
  howpublished = {EasyChair Preprint 466},
  year      = {EasyChair, 2018}}