MTSQL-R1: Towards Long-Horizon Multi-Turn Text-to-SQL via Agentic Training

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Guo, Taicheng, Wang, Hai, Liu, ChaoChun, Golalikhani, Mohsen, Chen, Xin, Zhang, Xiangliang, Reddy, Chandan K.
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910141193912320
author Guo, Taicheng
Wang, Hai
Liu, ChaoChun
Golalikhani, Mohsen
Chen, Xin
Zhang, Xiangliang
Reddy, Chandan K.
author_facet Guo, Taicheng
Wang, Hai
Liu, ChaoChun
Golalikhani, Mohsen
Chen, Xin
Zhang, Xiangliang
Reddy, Chandan K.
contents Multi-turn Text-to-SQL aims to translate a user's conversational utterances into executable SQL while preserving dialogue coherence and grounding to the target schema. However, most existing systems only regard this task as a simple text translation task and follow a short-horizon paradigm, generating a query per turn without execution, explicit verification, and refinement, which leads to non-executable or incoherent outputs. We present MTSQL-R1, an agentic training framework for long-horizon multi-turn Text-to-SQL. We cast the task as a Markov Decision Process (MDP) in which an agent interacts with (i) a database for execution feedback and (ii) a persistent dialogue memory for coherence verification, performing an iterative propose to execute -> verify -> refine cycle until all checks pass. Experiments on COSQL and SPARC demonstrate that MTSQL-R1 consistently outperforms strong baselines, highlighting the importance of environment-driven verification and memory-guided refinement for conversational semantic parsing. Full recipes (including code, trained models, logs, reasoning trajectories, etc.) will be released after the internal review to contribute to community research.
format Preprint
id arxiv_https___arxiv_org_abs_2510_12831
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle MTSQL-R1: Towards Long-Horizon Multi-Turn Text-to-SQL via Agentic Training
Guo, Taicheng
Wang, Hai
Liu, ChaoChun
Golalikhani, Mohsen
Chen, Xin
Zhang, Xiangliang
Reddy, Chandan K.
Computation and Language
Artificial Intelligence
Databases
Machine Learning
Multi-turn Text-to-SQL aims to translate a user's conversational utterances into executable SQL while preserving dialogue coherence and grounding to the target schema. However, most existing systems only regard this task as a simple text translation task and follow a short-horizon paradigm, generating a query per turn without execution, explicit verification, and refinement, which leads to non-executable or incoherent outputs. We present MTSQL-R1, an agentic training framework for long-horizon multi-turn Text-to-SQL. We cast the task as a Markov Decision Process (MDP) in which an agent interacts with (i) a database for execution feedback and (ii) a persistent dialogue memory for coherence verification, performing an iterative propose to execute -> verify -> refine cycle until all checks pass. Experiments on COSQL and SPARC demonstrate that MTSQL-R1 consistently outperforms strong baselines, highlighting the importance of environment-driven verification and memory-guided refinement for conversational semantic parsing. Full recipes (including code, trained models, logs, reasoning trajectories, etc.) will be released after the internal review to contribute to community research.
title MTSQL-R1: Towards Long-Horizon Multi-Turn Text-to-SQL via Agentic Training
topic Computation and Language
Artificial Intelligence
Databases
Machine Learning
url https://arxiv.org/abs/2510.12831