Learning to Infer Program Sketches
This addresses the challenge of program synthesis for users who need to generate code from intuitive inputs, representing an incremental advance in integrating neural and symbolic methods.
The paper tackles the problem of automatically writing code from human-friendly specifications like examples and natural language by combining pattern recognition and explicit reasoning, achieving state-of-the-art performance on a dataset of English description-to-code tasks.
Our goal is to build systems which write code automatically from the kinds of specifications humans can most easily provide, such as examples and natural language instruction. The key idea of this work is that a flexible combination of pattern recognition and explicit reasoning can be used to solve these complex programming problems. We propose a method for dynamically integrating these types of information. Our novel intermediate representation and training algorithm allow a program synthesis system to learn, without direct supervision, when to rely on pattern recognition and when to perform symbolic search. Our model matches the memorization and generalization performance of neural synthesis and symbolic search, respectively, and achieves state-of-the-art performance on a dataset of simple English description-to-code programming problems.