Text-to-SQL Task-oriented Dialogue Ontology Construction

2025-07-31

AI Summary

This work addresses the challenges of ontology construction in task-oriented dialogue (TOD) systems—namely, heavy reliance on manual annotation, limited interpretability, and poor controllability. We propose TeQoDO, the first unsupervised method that requires no human annotations or supervised training. TeQoDO synergistically integrates large language models’ (LLMs) text-to-SQL generation capability with dialogue-theoretic prompting to autonomously induce structured, interpretable, and controllable domain ontologies directly from raw dialogue data. The resulting ontology supports dynamic extension and demonstrates substantial improvements over transfer-learning baselines on downstream tasks such as dialogue state tracking. Extensive experiments on Wikipedia and ArXiv datasets validate TeQoDO’s effectiveness and scalability. Our approach establishes a novel paradigm for controllable, LLM-augmented TOD systems grounded in principled, theory-informed ontology induction.

Technology Category

• Natural Language Processing: Conversational AI/Dialog Systems
• Knowledge Representation and Reasoning: Ontologies
• Data Mining & Knowledge Management: Conversational Systems for Recommendation & Retrieval

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Search and Retrieval-Augmented AI: Search Tool Learning with LLM: Teaching LLMs to invoke search and make use of retrieved information
• Economics, Online Markets and Human Computation: Humans versus LLMs for data annotation and labeling

Abstract

Large language models (LLMs) are widely used as general-purpose knowledge sources, but they rely on parametric knowledge, limiting explainability and trustworthiness. In task-oriented dialogue (TOD) systems, this separation is explicit, using an external database structured by an explicit ontology to ensure explainability and controllability. However, building such ontologies requires manual labels or supervised training. We introduce TeQoDO: a Text-to-SQL task-oriented Dialogue Ontology construction method. Here, an LLM autonomously builds a TOD ontology from scratch without supervision using its inherent SQL programming capabilities combined with dialogue theory provided in the prompt. We show that TeQoDO outperforms transfer learning approaches, and its constructed ontology is competitive on a downstream dialogue state tracking task. Ablation studies demonstrate the key role of dialogue theory. TeQoDO also scales to allow construction of much larger ontologies, which we investigate on a Wikipedia and ArXiv dataset. We view this as a step towards broader application of ontologies to increase LLM explainability.

Problem

• Autonomous ontology construction without manual supervision
• Enhancing explainability in task-oriented dialogue systems
• Scaling ontology construction for large datasets

Innovation

• LLM autonomously builds TOD ontology
• Combines SQL programming with dialogue theory
• Scales to construct larger ontologies