A meta-contrastive learning with data augmentation framework for zero-shot stance detection

Published: 2024, Last Modified: 03 Aug 2025Expert Syst. Appl. 2024EveryoneRevisionsBibTeXCC BY-SA 4.0
Abstract: Highlights•Considering data scarcity and model generalizability for zero-shot stance detection.•Employing a generation model to produce target keyphrases for data augmentation.•Utilizing meta-contrastive learning to enhance the model’s generalizability.•Achieving state-of-the-art performance on multiple datasets.
Loading