We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:


Current browse context:


Change to browse by:


References & Citations

DBLP - CS Bibliography


(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Computer Science > Computation and Language

Title: Semantic Enhanced Text-to-SQL Parsing via Iteratively Learning Schema Linking Graph

Abstract: The generalizability to new databases is of vital importance to Text-to-SQL systems which aim to parse human utterances into SQL statements. Existing works achieve this goal by leveraging the exact matching method to identify the lexical matching between the question words and the schema items. However, these methods fail in other challenging scenarios, such as the synonym substitution in which the surface form differs between the corresponding question words and schema items. In this paper, we propose a framework named ISESL-SQL to iteratively build a semantic enhanced schema-linking graph between question tokens and database schemas. First, we extract a schema linking graph from PLMs through a probing procedure in an unsupervised manner. Then the schema linking graph is further optimized during the training process through a deep graph learning method. Meanwhile, we also design an auxiliary task called graph regularization to improve the schema information mentioned in the schema-linking graph. Extensive experiments on three benchmarks demonstrate that ISESL-SQL could consistently outperform the baselines and further investigations show its generalizability and robustness.
Comments: 8 pages, 3 figures, SIGKDD2022
Subjects: Computation and Language (cs.CL)
MSC classes: 03B65
ACM classes: I.7.0
Cite as: arXiv:2208.03903 [cs.CL]
  (or arXiv:2208.03903v1 [cs.CL] for this version)

Submission history

From: Aiwei Liu [view email]
[v1] Mon, 8 Aug 2022 03:59:33 GMT (3150kb,D)

Link back to: arXiv, form interface, contact.