Learning-to-Rank

We propose a unified generate-then-rank framework that can be flexibly incorporated with existing NLIDBs to consistently improve their translation accuracy.

We propose a generate-and-rank approach for accurate natural language to SQL translation.