TL-GRN: triplet loss-grouped reversible network

Shaolin Lv; Bin Sheng

doi:10.1117/12.3032831

5 July 2024 TL-GRN: triplet loss-grouped reversible network

Shaolin Lv, Bin Sheng

Proceedings Volume 13184, Third International Conference on Electronic Information Engineering and Data Processing (EIEDP 2024); 1318453 (2024) https://doi.org/10.1117/12.3032831
Event: 3rd International Conference on Electronic Information Engineering and Data Processing (EIEDP 2024), 2024, Kuala Lumpur, Malaysia

Abstract

This paper focuses on the problem of node classification on text-attributed graphs (TAGs), for which many techniques and methods have been proposed. Early Graph Neural Network methods often face many challenges when dealing with textual node classification tasks, such as incomplete, inaccurate and unbalanced data. These problems can lead to over-smoothing of the model during training, which in turn affects the ability of the model to capture local features and edge cases, thus reducing the accuracy of prediction and classification. In order to improve the performance of Graph Neural Networks in the textual node classification task, this paper explores a series of improvement and extension strategies, including reversible connection, residual connection, grouped convolution, and loss function design. These strategies are integrated into the GNN to form the model TL-GRN (Triplet Loss-Grouped Reversible Network) in this paper. By combining the advantages of Reversible connection and grouped convolution, we proposed a Grouped Reversible Network (GRN). This network structure aims to reduce the memory requirements of the model through reversibility while improving the representation of graph-structured data through grouped convolution. In addition, a Triplet Loss (TL) component is introduced in the design of the loss function, which is specifically used to define the similarity between groups of nodes of different classes. Experimental results on the ogbn-arxiv dataset show that TL-GRN is not only effective but also outperforms existing baseline models when dealing with textual node classification tasks.

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

Citation Download Citation

Shaolin Lv and Bin Sheng "TL-GRN: triplet loss-grouped reversible network", Proc. SPIE 13184, Third International Conference on Electronic Information Engineering and Data Processing (EIEDP 2024), 1318453 (5 July 2024); https://doi.org/10.1117/12.3032831

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

Members: $17.00

Non-members: $21.00 ADD TO CART

PROCEEDINGS
7 PAGES

DOWNLOAD PAPER SAVE TO MY LIBRARY

GET CITATION

RIGHTS & PERMISSIONS

Get copyright permission Get copyright permission on Copyright Marketplace

KEYWORDS

Performance modeling

Education and training

Data modeling

Convolution

Neural networks

Design

Lens design

Show All Keywords

Keywords/Phrases

Search In:

Publication Years