Neural Attribute Machines for Program Generation

TL;DR

Neural Attribute Machines integrate a logical grammar model with neural networks to improve program generation accuracy by enforcing grammatical constraints during training and generation.

Contribution

This paper introduces Neural Attribute Machines, a novel approach combining logical grammar models with neural networks to better learn and generate structured sequences.

Findings

NAMs significantly reduce grammar violations during generation

NAMs outperform traditional RNNs trained only on language samples

Incorporating grammar constraints improves sequence validity

Abstract

Recurrent neural networks have achieved remarkable success at generating sequences with complex structures, thanks to advances that include richer embeddings of input and cures for vanishing gradients. Trained only on sequences from a known grammar, though, they can still struggle to learn rules and constraints of the grammar. Neural Attribute Machines (NAMs) are equipped with a logical machine that represents the underlying grammar, which is used to teach the constraints to the neural machine by (i) augmenting the input sequence, and (ii) optimizing a custom loss function. Unlike traditional RNNs, NAMs are exposed to the grammar, as well as samples from the language of the grammar. During generation, NAMs make significantly fewer violations of the constraints of the underlying grammar than RNNs trained only on samples from the language of the grammar.