Program Synthesis and Semantic Parsing with Learned Code Idioms
This addresses the challenge of synthesizing general-purpose code from natural language for developers, though it is incremental as it builds on existing neural synthesis methods.
The paper tackles program synthesis from natural language by enabling neural synthesizers to interleave high-level and low-level reasoning using automatically mined code idioms, resulting in improved accuracy on semantic parsing datasets.
Program synthesis of general-purpose source code from natural language specifications is challenging due to the need to reason about high-level patterns in the target program and low-level implementation details at the same time. In this work, we present PATOIS, a system that allows a neural program synthesizer to explicitly interleave high-level and low-level reasoning at every generation step. It accomplishes this by automatically mining common code idioms from a given corpus, incorporating them into the underlying language for neural synthesis, and training a tree-based neural synthesizer to use these idioms during code generation. We evaluate PATOIS on two complex semantic parsing datasets and show that using learned code idioms improves the synthesizer's accuracy.