Dynamic Semantic Matching and Aggregation Network for Few-shot Intent Detection

TL;DR

This paper introduces a dynamic semantic matching network for few-shot intent detection, leveraging multi-head self-attention and a new evaluation setting to improve generalization to both seen and unseen classes.

Contribution

It proposes a novel semantic matching and aggregation network with dynamic regularization, enhancing high-level semantic component extraction for better few-shot intent detection.

Findings

Outperforms existing methods on few-shot intent detection tasks.

Effective in generalized settings with both seen and unseen classes.

Code and data are publicly available for reproducibility.

Abstract

Few-shot Intent Detection is challenging due to the scarcity of available annotated utterances. Although recent works demonstrate that multi-level matching plays an important role in transferring learned knowledge from seen training classes to novel testing classes, they rely on a static similarity measure and overly fine-grained matching components. These limitations inhibit generalizing capability towards Generalized Few-shot Learning settings where both seen and novel classes are co-existent. In this paper, we propose a novel Semantic Matching and Aggregation Network where semantic components are distilled from utterances via multi-head self-attention with additional dynamic regularization constraints. These semantic components capture high-level information, resulting in more effective matching between instances. Our multi-perspective matching method provides a comprehensive matching measure to enhance representations of both labeled and unlabeled instances. We also propose a more challenging evaluation setting that considers classification on the joint all-class label space. Extensive experimental results demonstrate the effectiveness of our method. Our code and data are publicly available.