Towards LLM-driven Dialogue State Tracking

TL;DR

This paper evaluates ChatGPT's effectiveness in dialogue state tracking and introduces LDST, an open-source framework that matches ChatGPT's performance using domain-slot instruction tuning on smaller models.

Contribution

It provides the first assessment of ChatGPT for DST and proposes LDST, a novel open-source approach that achieves comparable results with fewer resources.

Findings

ChatGPT performs exceptionally well in DST tasks.

LDST matches ChatGPT's performance using smaller models.

LDST shows significant improvements in zero-shot and few-shot settings.

Abstract

Dialogue State Tracking (DST) is of paramount importance in ensuring accurate tracking of user goals and system actions within task-oriented dialogue systems. The emergence of large language models (LLMs) such as GPT3 and ChatGPT has sparked considerable interest in assessing their efficacy across diverse applications. In this study, we conduct an initial examination of ChatGPT's capabilities in DST. Our evaluation uncovers the exceptional performance of ChatGPT in this task, offering valuable insights to researchers regarding its capabilities and providing useful directions for designing and enhancing dialogue systems. Despite its impressive performance, ChatGPT has significant limitations including its closed-source nature, request restrictions, raising data privacy concerns, and lacking local deployment capabilities. To address these concerns, we present LDST, an LLM-driven DST framework based on smaller, open-source foundation models. By utilizing a novel domain-slot instruction tuning method, LDST achieves performance on par with ChatGPT. Comprehensive evaluations across three distinct experimental settings, we find that LDST exhibits remarkable performance improvements in both zero-shot and few-shot setting compared to previous SOTA methods. The source code is provided for reproducibility.