Overview

Text-to-SQL studies how to turn natural-language questions into database queries that retrieve the intended information. Systems link questions to tables, columns, and values, then generate SQL using language models, structured intermediate plans, retrieval, or schema-aware semantic metadata. Common approaches add query decomposition, execution feedback, verification, and iterative repair; training methods include fine-tuning on curated or synthetic examples and reinforcement learning with execution-based rewards. Benchmarks such as Spider and BIRD test execution accuracy, but real deployments also expose challenges in large or changing schemas, ambiguity, multiple SQL dialects, and balancing reliability against cost and latency.

Latest papers

CardsList

186 papers20 papers in the last four weeks, down 9% on the four weeks before. 0.1% of all new papers.

  1. IESR:Efficient MCTS-Based Modular Reasoning for Text-to-SQL with Large Language Models

    Feb 5, 2026Tao Liu, Jiafan Lu, Bohan Yu +11Small Language ModelsText-to-SQL

  2. CORE-T: COherent REtrieval of Tables for Text-to-SQL

    Jan 19, 2026Hassan Soliman, Vivek Gupta, Dan Roth +1Text-to-SQLInformation Retrieval

  3. Effective and Efficient Threat Hunting with Small Language Models

    Dec 7, 2025Saleha Muzammil, Rahul Reddy, Vishal Kamalakrishnan +2CybersecuritySmall Language Models

  4. LearNAT: Learning NL2SQL with AST-guided Task Decomposition for Large Language Models

    Apr 3, 2025Weibin Liao, Xin Gao, Tianyu Jia +6LLM PlanningRL for Language Model Reasoning

  5. SQL-o1: Automatic Text-to-SQL Program Synthesis via Self-Reward Heuristic Search

    Feb 17, 2025Shuai Lyu, Haoran Luo, Ripeng Li +3Text-to-SQLLLM-Based Program Synthesis

  6. QDA-SQL: Questions Enhanced Dialogue Augmentation for Multi-Turn Text-to-SQL

    Jun 15, 2024Yinggang Sun, Ziming Guo, Haining Yu +5Synthetic Data GenerationQuestion Answering