A Framework to Generate High-Quality Datapoints for Multiple Novel Intent Detection

TL;DR

This paper introduces MNID, a cluster-based framework designed to efficiently detect and annotate multiple emerging intents in voice-command systems, improving accuracy with limited manual annotation budgets.

Contribution

The paper presents a novel cluster-based framework, MNID, for detecting and annotating multiple new intents simultaneously with minimal manual labeling effort.

Findings

MNID outperforms baseline methods in accuracy and F1-score.

Empirical validation on various benchmark datasets.

Efficient use of annotation budget improves intent detection.

Abstract

Systems like Voice-command based conversational agents are characterized by a pre-defined set of skills or intents to perform user specified tasks. In the course of time, newer intents may emerge requiring retraining. However, the newer intents may not be explicitly announced and need to be inferred dynamically. Thus, there are two important tasks at hand (a). identifying emerging new intents, (b). annotating data of the new intents so that the underlying classifier can be retrained efficiently. The tasks become specially challenging when a large number of new intents emerge simultaneously and there is a limited budget of manual annotation. In this paper, we propose MNID (Multiple Novel Intent Detection) which is a cluster based framework to detect multiple novel intents with budgeted human annotation cost. Empirical results on various benchmark datasets (of different sizes) demonstrate that MNID, by intelligently using the budget for annotation, outperforms the baseline methods in terms of accuracy and F1-score.