GraphCBAL-Sys: A Class-Balanced Active Learning System for Graphs

  • Chengcheng Yu
  • , Wenqian Zhou
  • , Jiahui Wang
  • , Fangshu Chen
  • , Jiapeng Zhu
  • , Xiang Li*
  • *Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

Active learning for Graph Neural Networks (GNNs) aims to select valuable unlabeled samples for annotation with a limited budget to maximize the GNNs’ performance at a low cost. However, most methods often result in imbalanced class distributions, leading to a bias toward majority classes, which undermines minority class performance and overall model effectiveness. To tackle this issue, we develop the Class-Balanced Active Learning System for Graphs GraphCBAL-Sys. It learns an optimal policy through reinforcement learning to acquire class-balanced and informative nodes for annotation. Additionally, GraphCBAL-Sys is capable of visualizing the internal processes and results during our model’s training and testing phases. Our demonstration video can be found here: https://b23.tv/yCLOIPw.

Original languageEnglish
Title of host publicationDatabase Systems for Advanced Applications - 30th International Conference, DASFAA 2025, Proceedings
EditorsFeida Zhu, Ee-Peng Lim, Philip S. Yu, Akiyo Nadamoto, Kyuseok Shim, Wei Ding, Bingxue Zhang
PublisherSpringer Science and Business Media Deutschland GmbH
Pages473-476
Number of pages4
ISBN (Print)9789819541577
DOIs
StatePublished - 2026
Event30th International Conference on Database Systems for Advanced Applications, DASFAA 2025 - Singapore, Singapore
Duration: 26 May 202529 May 2025

Publication series

NameLecture Notes in Computer Science
Volume15991 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference30th International Conference on Database Systems for Advanced Applications, DASFAA 2025
Country/TerritorySingapore
CitySingapore
Period26/05/2529/05/25

Keywords

  • Class balance
  • Graph Active learning
  • Reinforcement learning
  • Visualization

Fingerprint

Dive into the research topics of 'GraphCBAL-Sys: A Class-Balanced Active Learning System for Graphs'. Together they form a unique fingerprint.

Cite this