Exploring Zero and Few-shot Techniques for Intent Classifcation

Abstract

Conversational NLU providers often need to scale to thousands of intent-classifcation models where new customers often face the cold-start problem. Scaling to so many customers puts a constraint on storage space as well. In this paper, we explore four different zero and few-shot intent classifcation approaches with this low-resource constraint: 1) domain adaptation, 2) data augmentation, 3) zero-shot intent classifcation using descriptions large language models (LLMs), and 4) parameter-effcient fne-tuning of instruction-fnetuned language models. Our results show that all these approaches are effective to different degrees in low-resource settings. Parameter-effcient fne-tuning using T-few recipe (Liu et al., 2022) on Flan-T5 (Chung et al., 2022) yields the best performance even with just one sample per intent. We also show that the zero-shot method of prompting LLMs using intent descriptions is also very competitive.