Conformal Intent Classification and Clarification for Fast and Accurate Intent Recognition

TL;DR

In a comparative evaluation using seven intent recognition datasets, it is found that CICC generates small clarification questions and is capable of out-of-scope detection.

Abstract

We present Conformal Intent Classification and Clarification (CICC), a framework for fast and accurate intent classification for task-oriented dialogue systems. The framework turns heuristic uncertainty scores of any intent classifier into a clarification question that is guaranteed to contain the true intent at a pre-defined confidence level. By disambiguating between a small number of likely intents, the user query can be resolved quickly and accurately. Additionally, we propose to augment the framework for out-of-scope detection. In a comparative evaluation using seven intent recognition datasets we find that CICC generates small clarification questions and is capable of out-of-scope detection. CICC can help practitioners and researchers substantially in improving the user experience of dialogue agents with specific clarification questions.

Figures (9)

• Figure 1: The conformal intent classification and clarification interaction loop.
• Figure 2: Test set results for varying error rate $\alpha$.
• Figure 3: Intent distribution in ACID data set.
• Figure 4: Intent distribution in ATIS data set.
• Figure 5: Intent distribution in B77 data set.