Skip to search boxSkip to navigationSkip to main content

Injecting commonsense knowledge into prompt learning for zero-shot text classification

  • Jing Qian
    ,
  • Qi Chen
    ,
  • Yong Yue
    ,
  • Katie Atkinson
    ,
  • Gangmin Li
  • Xi'an Jiaotong-Liverpool University
    ,
  • University of Liverpool
Research Output: Chapter in Book/Report/Conference proceeding Conference contribution Peer-review

Abstract

The combination of pre-training and fine-tuning has become a default solution to Natural Language Processing (NLP) tasks. The emergence of prompt learning breaks such routine, especially in the scenarios of low data resources. Insufficient labelled data or even unseen classes are frequent problems in text classification, equipping Pre-trained Language Models (PLMs) with task-specific prompts helps get rid of the dilemma. However, general PLMs are barely provided with commonsense knowledge. In this work, we propose a KG-driven verbalizer that leverages commonsense Knowledge Graph (KG) to map label words with predefined classes. Specifically, we transform the mapping relationships into semantic relevance in the commonsense-injected embedding space. For zero-shot text classification task, experimental results exhibit the effectiveness of our KG-driven verbalizer on a Twitter dataset for natural disasters (i.e. HumAID) compared with other baselines.

Publication Information

Output type

Research Output: Chapter in Book/Report/Conference proceeding Conference contribution Peer-review

Original language

English

Pages from-to (Number of pages)

Pages 427-432 (6 pages)

Publication milestones

  • Published - 07/09/2023

Publication status

Published - 07/09/2023

Publisher

Association for Computing Machinery, United States

Publication series

  • Publication series name: ACM International Conference Proceeding Series

ISBN (Print)

9781450398411

ISBN (Electronic)

9781450398411

External Publication IDs

  • handle.net: 10547/626887
  • Scopus: 85173894088

Host publication title

ICMLC '23: Proceedings of the 2023 15th International Conference on Machine Learning and Computing

Access to documents

Publication metrics