ExeDec: Execution Decomposition for Compositional Generalization in Neural Program Synthesis

TL;DR

This paper introduces ExeDec, a decomposition-based approach for neural program synthesis that enhances compositional generalization by predicting execution subgoals, outperforming baselines and improving few-shot learning in large language models.

Contribution

The paper proposes ExeDec, a novel execution decomposition strategy that improves compositional generalization in neural program synthesis and demonstrates its effectiveness on benchmark datasets.

Findings

ExeDec improves synthesis performance over baselines.

ExeDec enhances compositional generalization in neural models.

ExeDec-style prompting benefits large language models in few-shot settings.

Abstract

When writing programs, people have the ability to tackle a new complex task by decomposing it into smaller and more familiar subtasks. While it is difficult to measure whether neural program synthesis methods have similar capabilities, we can measure whether they compositionally generalize, that is, whether a model that has been trained on the simpler subtasks is subsequently able to solve more complex tasks. In this paper, we characterize several different forms of compositional generalization that are desirable in program synthesis, forming a meta-benchmark which we use to create generalization tasks for two popular datasets, RobustFill and DeepCoder. We then propose ExeDec, a novel decomposition-based synthesis strategy that predicts execution subgoals to solve problems step-by-step informed by program execution at each step. When used with Transformer models trained from scratch, ExeDec has better synthesis performance and greatly improved compositional generalization ability compared to baselines. Finally, we use our benchmarks to demonstrate that LLMs struggle to compositionally generalize when asked to do programming-by-example in a few-shot setting, but an ExeDec-style prompting approach can improve the generalization ability and overall performance.