English

Learning Class-Transductive Intent Representations for Zero-shot Intent Detection

Computation and Language 2021-06-10 v2 Machine Learning

Abstract

Zero-shot intent detection (ZSID) aims to deal with the continuously emerging intents without annotated training data. However, existing ZSID systems suffer from two limitations: 1) They are not good at modeling the relationship between seen and unseen intents. 2) They cannot effectively recognize unseen intents under the generalized intent detection (GZSID) setting. A critical problem behind these limitations is that the representations of unseen intents cannot be learned in the training stage. To address this problem, we propose a novel framework that utilizes unseen class labels to learn Class-Transductive Intent Representations (CTIR). Specifically, we allow the model to predict unseen intents during training, with the corresponding label names serving as input utterances. On this basis, we introduce a multi-task learning objective, which encourages the model to learn the distinctions among intents, and a similarity scorer, which estimates the connections among intents more accurately. CTIR is easy to implement and can be integrated with existing methods. Experiments on two real-world datasets show that CTIR brings considerable improvement to the baseline systems.

Keywords

Cite

@article{arxiv.2012.01721,
  title  = {Learning Class-Transductive Intent Representations for Zero-shot Intent Detection},
  author = {Qingyi Si and Yuanxin Liu and Peng Fu and Zheng Lin and Jiangnan Li and Weiping Wang},
  journal= {arXiv preprint arXiv:2012.01721},
  year   = {2021}
}

Comments

IJCAI-2021

R2 v1 2026-06-23T20:41:44.315Z