Reproduce “Towards Robustness of Text-to-SQL Models against Synonym Substitution” on OpenResearch

Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.

Universal app for Apple silicon and Intel. Requires macOS 11 or later.

About this paper

Yujian Gan, Xinyun Chen, Qiuping Huang, Matthew Purver, John R. Woodward, Jinxia Xie, Pengsheng Huang

The authors introduce schema aliases for text-to-SQL, letting question synonyms serve as schema labels without retraining. RAT-SQL with BERT embeddings scores 69.7% exact-match accuracy on Spider development questions and 48.2% on Spider-Syn development questions. ManualMAS reaches 62.6% on Spider-Syn using its own substitutions, so generalization to unseen synonyms is not established.

Summary by alphaXiv.

Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.