MIGA: A Unified Multi-Task Generation Framework for Conversational Text-to-SQL

Conversational text-to-SQL is designed to translate multiturn natural language questions into their corresponding SQL queries. Most state-of-the-art conversational textto-SQL methods are incompatible with generative pretrained language models (PLMs), such as T5. In this paper, we present a two-stage unified MultI-task Generation frAmework (MIGA) that leverages PLMs' ability to tackle conversational text-to-SQL. In the pre-training stage, MIGA first decomposes the main task into several related subtasks and then unifies them into the same sequence-tosequence (Seq2Seq) paradigm with task-specific natural language prompts to boost the main task from multi-task training. Later in the fine-tuning stage, we propose four SQL perturbations to alleviate the error propagation problem. MIGA tends to achieve state-of-the-art performance on two benchmarks (SparC and CoSQL). We also provide extensive analyses and discussions to shed light on some new perspectives for conversational text-to-SQL.