Efficient Out-of-Scope Detection in Dialogue Systems via Uncertainty-Driven LLM Routing

TL;DR

This paper introduces a modular framework combining uncertainty estimation and fine-tuned large language models to improve out-of-scope intent detection in dialogue systems, achieving state-of-the-art results efficiently.

Contribution

It presents a novel approach that integrates uncertainty modeling with LLMs for more accurate and efficient out-of-scope detection in real-world dialogue systems.

Findings

State-of-the-art OOS detection performance on benchmark datasets

Effective balance between computational efficiency and accuracy

Successful deployment in a real-world dialogue system

Abstract

Out-of-scope (OOS) intent detection is a critical challenge in task-oriented dialogue systems (TODS), as it ensures robustness to unseen and ambiguous queries. In this work, we propose a novel but simple modular framework that combines uncertainty modeling with fine-tuned large language models (LLMs) for efficient and accurate OOS detection. The first step applies uncertainty estimation to the output of an in-scope intent detection classifier, which is currently deployed in a real-world TODS handling tens of thousands of user interactions daily. The second step then leverages an emerging LLM-based approach, where a fine-tuned LLM is triggered to make a final decision on instances with high uncertainty. Unlike prior approaches, our method effectively balances computational efficiency and performance, combining traditional approaches with LLMs and yielding state-of-the-art results on key OOS detection benchmarks, including real-world OOS data acquired from a deployed TODS.