Speak to your parser: Interactive Text-to-SQL with natural language feedback

We introduce the task of semantic parse correction with natural language feedback. Given a natural language utterance, most semantic parsing systems pose the problem as a one-shot translation problem where the utterance is mapped to a corresponding logical form. In this paper, we investigate a more interactive scenario where humans can further interact with the system by providing free-form natural language feedback to correct the system when it generates an inaccurate interpretation of an initial utterance.
We focus on natural language to SQL systems and construct a dataset of utterance, incorrect SQL interpretations and natural language feedback. We show that incorporating such rich form of feedback significantly improves the overall semantic parsing accuracy while still retaining the  flexibility of natural language interaction.