A significant amount of the world's knowledge is stored in relational databases.
Existing state-of-the-art approaches rely on reinforcement learning to reward the decoder when it generates any of the equivalent serializations.
Second, we show that the current division of data into training and test sets measures robustness to variations in the way questions are asked, but only partially tests how well systems generalize to new queries; therefore, we propose a complementary dataset split for evaluation of future work.
We define a new complex and cross-domain semantic parsing and text-to-SQL task where different complex SQL queries and databases appear in train and test sets.
SOTA for Semantic Parsing on spider