PICARD: Parsing Incrementally for Constrained Auto-Regressive Decoding from Language Models

TL;DR

PICARD is a method that constrains auto-regressive language model decoding using incremental parsing, significantly improving the validity and performance of models generating structured outputs like SQL.

Contribution

It introduces a novel incremental parsing approach to constrain language model decoding, enhancing validity and state-of-the-art performance in text-to-SQL tasks.

Findings

PICARD improves validity of generated SQL code.

PICARD achieves state-of-the-art results on Spider and CoSQL.

Constrained decoding reduces invalid outputs significantly.

Abstract

Large pre-trained language models for textual data have an unconstrained output space; at each decoding step, they can produce any of 10,000s of sub-word tokens. When fine-tuned to target constrained formal languages like SQL, these models often generate invalid code, rendering it unusable. We propose PICARD (code and trained models available at https://github.com/ElementAI/picard), a method for constraining auto-regressive decoders of language models through incremental parsing. PICARD helps to find valid output sequences by rejecting inadmissible tokens at each decoding step. On the challenging Spider and CoSQL text-to-SQL translation tasks, we show that PICARD transforms fine-tuned T5 models with passable performance into state-of-the-art solutions.