Lune

AAAI2024Top-tier venue

Every Node Is Different: Dynamically Fusing Self-Supervised Tasks for Attributed Graph Clustering

Pengfei Zhu, Qian Wang, Yu Wang, Jialu Li, Qinghua Hu

Abstract

Attributed graph clustering is an unsupervised task that partitions nodes into different groups. Self-supervised learning (SSL) shows great potential in handling this task, and some recent studies simultaneously learn multiple SSL tasks to further boost performance. Currently, different SSL tasks are assigned the same set of weights for all graph nodes. However, we observe that some graph nodes whose neighbors are in different groups require significantly different emphases on SSL tasks. In this paper, we propose to dynamically learn the weights of SSL tasks for different nodes and fuse the embeddings learned from different SSL tasks to boost performance. We design an innovative graph clustering approach, namely Dynamically Fusing Self-Supervised Learning (DyFSS). Specifically, DyFSS fuses features extracted from diverse SSL tasks using distinct weights derived from a gating network. To effectively learn the gating network, we design a dual-level self-supervised strategy that incorporates pseudo labels and the graph structure. Extensive experiments on five datasets show that DyFSS outperforms the state-of-the-art multi-task SSL methods by up to 8.66% on the accuracy metric. The code of DyFSS is available at: https://github.com/q086/DyFSS .

Ask Lune

Know this paper in minutes, not an afternoon.

Lune can read every page of it and the top-tier papers that cite it, and gives you its claims, its evidence and its limits, each one cited to the paper it comes from. No skimming, no AI slop.

2 top-tier papers have cited it since. Know what they found before your reviewers do.

Questions to start from

Lune can draw on

the full paper · the top-tier papers that cite it · related work

Get the answer

Free to start. No credit card required.

Lineage

  • One Node One Model: Featuring the Missing-Half for Graph Clustering cites Structural Deep Clustering Network
  • Weighted Graph Clustering via Scale Contraction and Graph Structure Learning cites Structural Deep Clustering Network
  • From Canonical Correlation Analysis to Self-supervised Graph Neural Networks cites Graph Contrastive Learning with Adaptive Augmentation
  • Large-Scale Representation Learning on Graphs via Bootstrapping cites Graph Contrastive Learning with Adaptive Augmentation
  • GraphMAE: Self-Supervised Masked Graph Autoencoders cites From Canonical Correlation Analysis to Self-supervised Graph Neural Networks
  • GraphMAE: Self-Supervised Masked Graph Autoencoders cites Large-Scale Representation Learning on Graphs via Bootstrapping
  • GraphMAE: Self-Supervised Masked Graph Autoencoders cites Graph Contrastive Learning with Adaptive Augmentation
Builds onCited byLeft to right by year; the more cited, the higher.

Cited by top-tier papers2

  1. One Node One Model: Featuring the Missing-Half for Graph ClusteringXuanting Xie, Bingheng Li, Erlin Pan, Zhaochen Guo et al.AAAI 2025 · 4 citations
  2. Weighted Graph Clustering via Scale Contraction and Graph Structure LearningHaobing Liu, Yinuo Zhang, Tingting Wang, Ruobing Jiang et al.WWW 2026 · 4 citations

Builds on9

  1. Graph Contrastive Learning with Adaptive AugmentationYanqiao Zhu, Yichen Xu, Feng Yu, Qiang Liu et al.WWW 2021 · 1,415 citations
  2. Structural Deep Clustering NetworkDeyu Bo, Xiao Wang, Chuan Shi, Meiqi Zhu et al.WWW 2020 · 645 citations
  3. GraphMAE: Self-Supervised Masked Graph AutoencodersZhenyu Hou, Xiao Liu, Yukuo Cen, Yuxiao Dong et al.KDD 2022 · 533 citations
  4. From Canonical Correlation Analysis to Self-supervised Graph Neural NetworksHengrui Zhang, Qitian Wu, Junchi Yan, David Wipf et al.NeurIPS 2021 · 319 citations
  5. Large-Scale Representation Learning on Graphs via BootstrappingShantanu Thakoor, Corentin Tallec, Mohammad Gheshlaghi Azar, Mehdi Azabou et al.ICLR 2022 · 311 citations

Related papers

  1. Automated Self-Supervised Learning for GraphsWei Jin, Xiaorui Liu, Xiangyu Zhao, Yao Ma et al.ICLR 2022 · 95 citations
  2. Self-Supervised Graph Attention Networks for Deep Weighted Multi-View ClusteringZongmo Huang, Yazhou Ren, Xiaorong Pu, Shudong Huang et al.AAAI 2023 · 50 citations
  3. Multi-task Self-supervised Graph Neural Networks Enable Stronger Task GeneralizationMingxuan Ju, Tong Zhao, Qianlong Wen, Wenhao Yu et al.ICLR 2023 · 7 citations
  4. Deep Multi-view Graph Clustering via Attribute-aware Bidirectional Structural Refinement and Pseudo-label Guided Multi-level FusionYouqing Wang, Tianxiang Zhao, Mengyuan Xin, Ye Su et al.ICML 2026
  5. Deep Graph Clustering via Dual Correlation ReductionYue Liu, Wenxuan Tu, Sihang Zhou, Xinwang Liu et al.AAAI 2022 · 300 citations
Dusk over the sea between two cliffs drawn in fine vertical lines