DCMM-SQL: Automated Data-Centric Pipeline and Multi-Model Collaboration Training for Text-to-SQL Model

2025-10-27

AI Summary

To address low data quality and weak generalization of single models in Text-to-SQL tasks, this paper proposes the first data-centric solution. Our method introduces a fully automated, data-driven pipeline integrating adaptive data repair and error-guided augmentation, coupled with a multi-model collaborative iterative training framework that achieves joint optimization through diversified training, capability complementarity, and ensemble inference. Built upon an agent-based, multi-stage optimization architecture, the approach attains state-of-the-art performance (rank #1 on the leaderboard) using lightweight models under 70B parameters. Ablation studies quantitatively demonstrate substantial gains from both the data-centric pipeline and the multi-model collaboration mechanism. This work establishes a reproducible, scalable paradigm for co-optimizing data and models in Text-to-SQL, advancing beyond traditional model-centric approaches.

Technology Category

• Search and Optimization: Learning to Search
• Natural Language Processing: Language Grounding & Multi-modal NLP
• Machine Learning: Large Multimodal Models (LMMs)

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets

Abstract

Text-to-SQL tasks have gained attractive improvements since the release of ChatGPT. Among them, agent-based frameworks have been widely used in this field. However, the impact of data-centric strategies on text-to-SQL tasks has rarely been explored. In this paper, we systemically design a fully automated data-centric pipeline for text-to-SQL tasks, including emph{adaptive data repair}, which can automatically find and fix errors in the training dataset; and emph{error data augmentation}, where we specifically diffuse and enhance erroneous data predicted by the initially trained models. Meanwhile, we propose a Multi-Model collaboration training schema, aiming to train multiple models with different augmented data, enabling them to possess distinct capabilities and work together to complement each other, because it has been found that the capability of a single fine-tuned model is very limited. Furthermore, we utilize an ensemble strategy to integrate the capabilities of multiple models to solve a multiple-choice question, aiming to further improve the accuracy of text-to-SQL tasks. The experiment results and ablation study have demonstrated the effectiveness of data-centric pipeline and Multi-Model(MM) interactive iterative strategies, achieving first place in lightweight text-to-SQL models (within 70B).

Problem

• Automating data repair and augmentation for text-to-SQL training datasets
• Developing multi-model collaboration training with distinct specialized capabilities
• Improving accuracy of lightweight text-to-SQL models through ensemble strategies

Innovation

• Automated data-centric pipeline repairs and augments data
• Multi-model collaboration training with distinct capabilities
• Ensemble strategy integrates multiple models for accuracy