Geometer: Graph Few-Shot Class-Incremental Learning via Prototype Representation

被引:4
|
作者
Lu, Bin [1 ]
Gan, Xiaoying [1 ]
Yang, Lina [1 ]
Zhang, Weinan [1 ]
Fu, Luoyi [1 ]
Wang, Xinbing [1 ]
机构
[1] Shanghai Jiao Tong Univ, Shanghai, Peoples R China
关键词
Node Classification; Graph Neural Network; Few-Shot Learning; Class-Incremental Learning;
D O I
10.1145/3534678.3539280
中图分类号
TP [自动化技术、计算机技术];
学科分类号
0812 ;
摘要
With the tremendous expansion of graphs data, node classification shows its great importance in many real-world applications. Existing graph neural network based methods mainly focus on classifying unlabeled nodes within fixed classes with abundant labeling. However, in many practical scenarios, graph evolves with emergence of new nodes and edges. Novel classes appear incrementally along with few labeling due to its newly emergence or lack of exploration. In this paper, we focus on this challenging but practical graph few-shot class-incremental learning (GFSCIL) problem and propose a novel method called Geometer. Instead of replacing and retraining the fully connected neural network classifer, Geometer predicts the label of a node by finding the nearest class prototype. Prototype is a vector representing a class in the metric space. With the pop-up of novel classes, Geometer learns and adjusts the attention-based prototypes by observing the geometric proximity, uniformity and separability. Teacher-student knowledge distillation and biased sampling are further introduced to mitigate catastrophic forgetting and unbalanced labeling problem respectively. Experimental results on four public datasets demonstrate that Geometer achieves a substantial improvement of 9.46% to 27.60% over state-of-the-art methods.
引用
收藏
页码:1152 / 1161
页数:10
相关论文
共 50 条
  • [1] Graph Few-shot Class-incremental Learning
    Tan, Zhen
    Ding, Kaize
    Guo, Ruocheng
    Liu, Huan
    [J]. WSDM'22: PROCEEDINGS OF THE FIFTEENTH ACM INTERNATIONAL CONFERENCE ON WEB SEARCH AND DATA MINING, 2022, : 987 - 996
  • [2] Flexible few-shot class-incremental learning with prototype container
    Xu, Xinlei
    Wang, Zhe
    Fu, Zhiling
    Guo, Wei
    Chi, Ziqiu
    Li, Dongdong
    [J]. NEURAL COMPUTING & APPLICATIONS, 2023, 35 (15): : 10875 - 10889
  • [3] Few-shot class-incremental audio classification via discriminative prototype learning
    Xie, Wei
    Li, Yanxiong
    He, Qianhua
    Cao, Wenchang
    [J]. EXPERT SYSTEMS WITH APPLICATIONS, 2023, 225
  • [4] Few-shot class-incremental learning based on representation enhancement
    Yao, Guangle
    Zhu, Juntao
    Zhou, Wenlong
    Li, Jun
    [J]. JOURNAL OF ELECTRONIC IMAGING, 2022, 31 (04)
  • [5] Few-Shot Class-Incremental SAR Target Recognition via Cosine Prototype Learning
    Zhao, Yan
    Zhao, Lingjun
    Ding, Ding
    Hu, Dewen
    Kuang, Gangyao
    Liu, Li
    [J]. IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING, 2023, 61
  • [6] Few-Shot Class-Incremental Learning via Training-Free Prototype Calibration
    Wang, Qi-Wei
    Zhou, Da-Wei
    Zhang, Yi-Kai
    Zhan, De-Chuan
    Ye, Han-Jia
    [J]. ADVANCES IN NEURAL INFORMATION PROCESSING SYSTEMS 36 (NEURIPS 2023), 2023,
  • [7] Knowledge Representation by Generic Models for Few-Shot Class-Incremental Learning
    Chen, Xiaodong
    Jiang, Weijie
    Huang, Zhiyong
    Su, Jiangwen
    Yu, Yuanlong
    [J]. ADVANCES IN NATURAL COMPUTATION, FUZZY SYSTEMS AND KNOWLEDGE DISCOVERY, ICNC-FSKD 2022, 2023, 153 : 1237 - 1247
  • [8] A survey on few-shot class-incremental learning
    Tian, Songsong
    Li, Lusi
    Li, Weijun
    Ran, Hang
    Ning, Xin
    Tiwari, Prayag
    [J]. NEURAL NETWORKS, 2024, 169 : 307 - 324
  • [9] A survey on few-shot class-incremental learning
    Tian, Songsong
    Li, Lusi
    Li, Weijun
    Ran, Hang
    Ning, Xin
    Tiwari, Prayag
    [J]. Neural Networks, 2024, 169 : 307 - 324
  • [10] Constrained Few-shot Class-incremental Learning
    Hersche, Michael
    Karunaratne, Geethan
    Cherubini, Giovanni
    Benini, Luca
    Sebastian, Abu
    Rahimi, Abbas
    [J]. 2022 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR), 2022, : 9047 - 9057