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

Abstract

Systems like Voice-command based conversational agents are characterized by a predefined 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 F1score.

Intent Identification and Entity Extraction for Healthcare Queries in Indic Languages

“…(Mullick et al, 2022a) explore intent classification on legal data. People also work on different detection approaches -few shot (Xia et al, 2021), zero shot (Xia et al, 2018), clustering frameworks (Mullick et al, 2022b). (Yani et al, 2022;Sufi and Alsulami, 2021;Zhao et al, 2021) all explore entity detection tasks.…”