C3LPGCN:Integrating Contrastive Learning and Cooperative Learning with Prompt into Graph Convolutional Network for Aspect-based Sentiment Analysis

Ye He, Shihao Zou, YuzheChen YuzheChen, Xianying Huang


Abstract
Aspect-based Sentiment Analysis (ABSA) is a fine-grained task. Recently, using graph convolutional networks (GCNs) to model syntactic information has become a popular topic. In addition, a growing consensus exists to enhance sentence representation using contrastive learning. However, when modeling syntactic information, incorrect syntactic structure may introduce additional noise. Meanwhile, we believe that contrastive learning implicitly introduce label information as priori. Therefore, we propose C3LPGCN, which integrates Contrastive Learning and Cooperative Learning with Prompt into GCN. Specifically, to alleviate the noise when modeling syntactic information, we propose mask-aware aspect information filter, which combines prompt information of template with aspect information to filter the syntactic information. Besides, we propose prompt-based contrastive learning and cooperative learning to utilise the label information further. On the one hand, we construct prompts containing labels for contrastive learning, by which the model can focus more on task-relevant features. On the other hand, cooperative learning further extracts label information by aligning input samples’ representation and output distribution with label samples. Extensive experiments on three datasets demonstrate that our method significantly improves the model’s performance compared to traditional contrastive learning methods. Moreover, our C3LPGCN outperforms state-of-the-art methods. Our source code and final models are publicly available at github
Anthology ID:
2024.findings-naacl.205
Volume:
Findings of the Association for Computational Linguistics: NAACL 2024
Month:
June
Year:
2024
Address:
Mexico City, Mexico
Editors:
Kevin Duh, Helena Gomez, Steven Bethard
Venue:
Findings
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
3237–3247
Language:
URL:
https://aclanthology.org/2024.findings-naacl.205
DOI:
Bibkey:
Cite (ACL):
Ye He, Shihao Zou, YuzheChen YuzheChen, and Xianying Huang. 2024. C3LPGCN:Integrating Contrastive Learning and Cooperative Learning with Prompt into Graph Convolutional Network for Aspect-based Sentiment Analysis. In Findings of the Association for Computational Linguistics: NAACL 2024, pages 3237–3247, Mexico City, Mexico. Association for Computational Linguistics.
Cite (Informal):
C3LPGCN:Integrating Contrastive Learning and Cooperative Learning with Prompt into Graph Convolutional Network for Aspect-based Sentiment Analysis (He et al., Findings 2024)
Copy Citation:
PDF:
https://aclanthology.org/2024.findings-naacl.205.pdf
Copyright:
 2024.findings-naacl.205.copyright.pdf