CoDial: Interpretable Task-Oriented Dialogue Systems Through Dialogue Flow Alignment

TL;DR

CoDial introduces a novel, interpretable framework for task-oriented dialogue systems by converting schemas into programmatic code, achieving state-of-the-art results and enabling human-guided improvements.

Contribution

The paper presents a new approach converting task schemas into programmatic code for dialogue systems, enhancing interpretability and transferability across unseen tasks.

Findings

Achieves state-of-the-art performance on benchmark datasets.

Provides inherent interpretability through dialogue flow alignment.

Enables iterative human and LLM feedback for continuous improvement.

Abstract

Building Task-Oriented Dialogue (TOD) systems that generalize across different tasks remains a challenging problem. Data-driven approaches often struggle to transfer effectively to unseen tasks. While recent schema-based TOD frameworks improve generalization by decoupling task logic from language understanding, their reliance on neural or generative models often obscures how task schemas influence behaviour and hence impair interpretability. In this work, we introduce a novel framework, CoDial (Code for Dialogue), at the core of which is converting a predefined task schema to a structured heterogeneous graph and then to programmatic LLM guardrailing code, such as NVIDIA's Colang. The pipeline enables efficient and interpretable alignment of dialogue policies during inference. We introduce two paradigms for LLM guardrailing code generation, $\text{CoDial}_{\text{free}}$ and $\text{CoDial}_{\text{structured}}$, and propose a mechanism that integrates human feedback to iteratively improve the generated code. Empirically, CoDial achieves state-of-the-art (SOTA) performance on the widely used benchmark datasets, while providing inherent interpretability in the design. We additionally demonstrate CoDial's iterative improvement via manual and LLM-aided feedback, making it a practical tool for human-guided alignment of LLMs in unseen domains.