Using Database Rule for Weak Supervised Text-to-SQL Generation

Tong Guo, Huilin Gao
7/1/2019
cs.CL

Official code: guotong1988/weak-supervised-Rule-Text2SQL (12 stars, Apache-2.0, PyTorch, last commit 2021-05-11).

Abstract

We present a simple way to do the task of text-to-SQL problem with weak supervision. We call it Rule-SQL. Given the question and the answer from the database table without the SQL logic form, Rule-SQL use the rules based on table column names and question string for the SQL exploration first and then use the explored SQL for supervised training. We design several rules for reducing the exploration search space. For the deep model, we leverage BERT for the representation layer and separate the model to SELECT, AGG and WHERE parts. The experiment result on WikiSQL outperforms the strong baseline of full supervision and is comparable to the start-of-the-art weak supervised mothods.

View on arXivView PDF

Code Implementations(1)

guotong1988/weak-supervised-Rule-Text2SQLOfficial100%
PyTorchApache-2.0

Using Database Rule for Weak Supervised Text-to-SQL Generation https://arxiv.org/abs/1907.00620

122Python, Jupyter NotebookJul 25, 20195 years agoApache-2.0
deep-learningdeep-neural-networksnl2sqlpytorchsemantic-parsing+1 more

Similar Papers

Content Enhanced BERT-based Text-to-SQL Generation

82% match1 repo1952019

Graphix-T5: Mixing Pre-Trained Transformers with Graph-Aware Layers for Text-to-SQL Parsing

77% match2023

ProS: Prompting-to-Simulate Generalized Knowledge for Universal Cross-Domain Retrieval

77% match1 repo192023

Relevance-guided Supervision for OpenQA with ColBERT

77% match2 repos3.9k2020

DC-W2S: Dual-Consensus Weak-to-Strong Training for Reliable Process Reward Modeling in Biological Reasoning

76% match2026

Self-Study Reconsidered: The Hidden Fragility of Learning from Self-Generated QA

76% match2026

Cite this paper

@article{guo2019using,
  title  = {Using Database Rule for Weak Supervised Text-to-SQL Generation},
  author = {Tong Guo and Huilin Gao},
  year   = {2019},
  eprint = {1907.00620},
  archivePrefix = {arXiv},
  url    = {https://arxiv.org/abs/1907.00620}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/1907.00620.svg)](https://codeofpaper.com/paper/1907.00620)

Discussion