Toward Trustworthy Neural Program Synthesis

TL;DR

This paper presents a probabilistic approach to estimate the correctness of programs generated by large language models, improving trustworthiness and interpretability of program synthesis.

Contribution

It introduces a method that samples candidate programs and predicates to produce well-calibrated correctness probabilities and explainability, enhancing existing neural program synthesis techniques.

Findings

Achieves state-of-the-art accuracy in program generation

Provides well-calibrated correctness probabilities

Predicates improve human understanding of generated code

Abstract

We develop an approach to estimate the probability that a program sampled from a large language model is correct. Given a natural language description of a programming problem, our method samples both candidate programs as well as candidate predicates specifying how the program should behave. This allows learning a model that forms a well-calibrated probabilistic prediction of program correctness. Our system also infers which predicates are useful to explain the behavior of the generated code, and humans preferred these in a human study over raw language model outputs. Our method is simple, easy to implement, and maintains state of the art generation accuracy results.