Tracking Interaction States for Multi-Turn Text-to-SQL Semantic Parsing

The task of multi-turn text-to-SQL semantic parsing aims to translate natural language utterances in an interaction into SQL queries order answer them using a database which normally contains multiple table schemas. Previous studies on this usually utilized contextual information enrich utterance representations and further influence the decoding process. While they ignored describe track states are determined by history related with intent current utterance. In paper, two kinds defined based schema items keywords separately. A relational graph neural network non-linear layer designed update these respectively. dynamic schema-state SQL-state then decode query corresponding Experimental results challenging CoSQL dataset demonstrate effectiveness our proposed method, achieves better performance than other published methods leaderboard.