All Labels Together: Low-shot Intent Detection with an Efficient Label Semantic Encoding Paradigm

TL;DR

This paper introduces an end-to-end system that fully leverages label semantics for low-shot intent detection, achieving state-of-the-art results and enabling zero-shot cross-domain generalization through a novel pretraining strategy.

Contribution

It proposes a new One-to-All approach that compares utterances with all label candidates and a pretraining method using paraphrasing for zero-shot transfer.

Findings

State-of-the-art performance in 1-, 3-, and 5-shot settings

Effective zero-shot cross-domain generalization

Significant improvement over existing methods in low-resource scenarios

Abstract

In intent detection tasks, leveraging meaningful semantic information from intent labels can be particularly beneficial for few-shot scenarios. However, existing few-shot intent detection methods either ignore the intent labels, (e.g. treating intents as indices) or do not fully utilize this information (e.g. only using part of the intent labels). In this work, we present an end-to-end One-to-All system that enables the comparison of an input utterance with all label candidates. The system can then fully utilize label semantics in this way. Experiments on three few-shot intent detection tasks demonstrate that One-to-All is especially effective when the training resource is extremely scarce, achieving state-of-the-art performance in 1-, 3- and 5-shot settings. Moreover, we present a novel pretraining strategy for our model that utilizes indirect supervision from paraphrasing, enabling zero-shot cross-domain generalization on intent detection tasks. Our code is at https://github.com/jiangshdd/AllLablesTogether.