What functions can Graph Neural Networks compute on random graphs? The role of Positional Encoding

被引:0
|
作者
Keriven, Nicolas [1 ]
Vaiter, Samuel [2 ]
机构
[1] CNRS, IRISA, Rennes, France
[2] CNRS, LJAD, Nice, France
关键词
D O I
暂无
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
We aim to deepen the theoretical understanding of Graph Neural Networks (GNNs) on large graphs, with a focus on their expressive power. Existing analyses relate this notion to the graph isomorphism problem, which is mostly relevant for graphs of small sizes, or studied graph classification or regression tasks, while prediction tasks on nodes are far more relevant on large graphs. Recently, several works showed that, on very general random graphs models, GNNs converge to certains functions as the number of nodes grows. In this paper, we provide a more complete and intuitive description of the function space generated by equivariant GNNs for node-tasks, through general notions of convergence that encompass several previous examples. We emphasize the role of input node features, and study the impact of node Positional Encodings (PEs), a recent line of work that has been shown to yield state-of-the-art results in practice. Through the study of several examples of PEs on large random graphs, we extend previously known universality results to significantly more general models. Our theoretical results hint at some normalization tricks, which is shown numerically to have a positive impact on GNN generalization on synthetic and real data. Our proofs contain new concentration inequalities of independent interest.
引用
收藏
页数:27
相关论文
共 50 条
  • [41] What energy functions can be minimized via graph cuts?
    Kolmogorov, V
    Zabih, R
    COMPUTER VISION - ECCV 2002 PT III, 2002, 2352 : 65 - 81
  • [42] What energy functions can be minimized via graph cuts?
    Kolmogorov, V
    Zabih, R
    IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE, 2004, 26 (02) : 147 - 159
  • [43] Maneuver Prediction Using Traffic Scene Graphs via Graph Neural Networks and Recurrent Neural Networks
    Rama, Petrit
    Bajcinca, Naim
    INTERNATIONAL JOURNAL OF SEMANTIC COMPUTING, 2023, 17 (03) : 349 - 370
  • [44] Facilitating Graph Neural Networks with Random Walk on Simplicial Complexes
    Zhou, Cai
    Wang, Xiyuan
    Zhang, Muhan
    ADVANCES IN NEURAL INFORMATION PROCESSING SYSTEMS 36 (NEURIPS 2023), 2023,
  • [45] The Surprising Power of Graph Neural Networks with Random Node Initialization
    Abboud, Ralph
    Ceylan, Ismail Ilkan
    Grohe, Martin
    Lukasiewicz, Thomas
    PROCEEDINGS OF THE THIRTIETH INTERNATIONAL JOINT CONFERENCE ON ARTIFICIAL INTELLIGENCE, IJCAI 2021, 2021, : 2112 - 2118
  • [46] Graph Neural Networks with feature and structure aware random walk
    Zhuo, Wei
    Yu, Chenyun
    Tan, Guang
    arXiv, 2021,
  • [47] DropGNN: Random Dropouts Increase the Expressiveness of Graph Neural Networks
    Papp, Pal Andras
    Martinkus, Karolis
    Faber, Lukas
    Wattenhofer, Roger
    ADVANCES IN NEURAL INFORMATION PROCESSING SYSTEMS 34 (NEURIPS 2021), 2021, 34
  • [48] Conditional Random Field Enhanced Graph Convolutional Neural Networks
    Gao, Hongchang
    Pei, Jian
    Huang, Heng
    KDD'19: PROCEEDINGS OF THE 25TH ACM SIGKDD INTERNATIONAL CONFERENCCE ON KNOWLEDGE DISCOVERY AND DATA MINING, 2019, : 276 - 284
  • [49] Structural Temporal Graph Neural Networks for Anomaly Detection in Dynamic Graphs
    Cai, Lei
    Chen, Zhengzhang
    Luo, Chen
    Gui, Jiaping
    Ni, Jingchao
    Li, Ding
    PROCEEDINGS OF THE 30TH ACM INTERNATIONAL CONFERENCE ON INFORMATION & KNOWLEDGE MANAGEMENT, CIKM 2021, 2021, : 3747 - 3756
  • [50] Advancing Cybersecurity: Graph Neural Networks in Threat Intelligence Knowledge Graphs
    Li, Langsha
    Qiang, Feng
    Ma, Li
    PROCEEDINGS OF INTERNATIONAL CONFERENCE ON ALGORITHMS, SOFTWARE ENGINEERING, AND NETWORK SECURITY, ASENS 2024, 2024, : 737 - 741