Contrastive learning based on dual-path models

Dengdi Sun; Xiaoxin Zhang; Bin Luo; Zhuanlian Ding

doi:10.1117/12.3033038

5 July 2024 Contrastive learning based on dual-path models

Dengdi Sun, Xiaoxin Zhang, Bin Luo, Zhuanlian Ding

Proceedings Volume 13184, Third International Conference on Electronic Information Engineering and Data Processing (EIEDP 2024); 131846B (2024) https://doi.org/10.1117/12.3033038
Event: 3rd International Conference on Electronic Information Engineering and Data Processing (EIEDP 2024), 2024, Kuala Lumpur, Malaysia

Abstract

Unsupervised graph representation learning is a trending research topic, driven by the versatility of graph data. Graph neural networks (GNNs) and graph autoencoders (GAEs), alongside their derivatives like ARGA, MGAE, and GALA, have made significant strides. Yet, the practical challenge of effectively harnessing node features while maintaining graph structural integrity remains. This study introduces a novel dual-model path comparison approach for graph representation learning. Its dual-path encoder design crafts multiple high-level representations, capturing the graph's intricacies. We enhance these representations using reconstruction and cosine losses, ensuring their optimization and fusion. Our model notably boosts graph representation performance, offering streamlined, concise, and adaptable representations for downstream applications. This breakthrough holds considerable theoretical and practical weight in unsupervised graph learning, showcasing GNNs and GAEs' prowess and charting new solutions for balancing node features and graph structures. This work is poised to inform and inspire future research in the field.

(2024) Published by SPIE. Downloading of the abstract is permitted for personal use only.

Citation Download Citation

Dengdi Sun, Xiaoxin Zhang, Bin Luo, and Zhuanlian Ding "Contrastive learning based on dual-path models", Proc. SPIE 13184, Third International Conference on Electronic Information Engineering and Data Processing (EIEDP 2024), 131846B (5 July 2024); https://doi.org/10.1117/12.3033038

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available