Grounded Adaptation for Zero-shot Executable Semantic Parsing
This addresses the challenge of zero-shot adaptation for semantic parsing, enabling parsers to work in new domains without costly data annotation, though it is incremental as it builds on existing parser frameworks.
The paper tackles the problem of adapting semantic parsers to new environments like database schemas without labeled data, proposing GAZP which synthesizes and verifies examples to improve accuracy, achieving gains in logical form and execution accuracy on tasks like Spider, Sparc, and CoSQL.
We propose Grounded Adaptation for Zero-shot Executable Semantic Parsing (GAZP) to adapt an existing semantic parser to new environments (e.g. new database schemas). GAZP combines a forward semantic parser with a backward utterance generator to synthesize data (e.g. utterances and SQL queries) in the new environment, then selects cycle-consistent examples to adapt the parser. Unlike data-augmentation, which typically synthesizes unverified examples in the training environment, GAZP synthesizes examples in the new environment whose input-output consistency are verified. On the Spider, Sparc, and CoSQL zero-shot semantic parsing tasks, GAZP improves logical form and execution accuracy of the baseline parser. Our analyses show that GAZP outperforms data-augmentation in the training environment, performance increases with the amount of GAZP-synthesized data, and cycle-consistency is central to successful adaptation.