A Biased Graph Neural Network Sampler with Near-Optimal Regret

被引:0
作者
Zhang, Qingru [1 ]
Wipf, David [2 ]
Gan, Quan [2 ]
Song, Le [1 ,3 ]
机构
[1] Georgia Inst Technol, Atlanta, GA 30332 USA
[2] Amazon Shanghai AI Lab, Shanghai, Peoples R China
[3] Mohamed Bin Zayed Univ Artificial Intelligence, Abu Dhabi, U Arab Emirates
来源
ADVANCES IN NEURAL INFORMATION PROCESSING SYSTEMS 34 (NEURIPS 2021) | 2021年 / 34卷
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
Graph neural networks (GNN) have recently emerged as a vehicle for applying deep network architectures to graph and relational data. However, given the increasing size of industrial datasets, in many practical situations the message passing computations required for sharing information across GNN layers are no longer scalable. Although various sampling methods have been introduced to approximate full-graph training within a tractable budget, there remain unresolved complications such as high variances and limited theoretical guarantees. To address these issues, we build upon existing work and treat GNN neighbor sampling as a multi-armed bandit problem but with a newly-designed reward function that introduces some degree of bias designed to reduce variance and avoid unstable, possibly-unbounded pay outs. And unlike prior bandit-GNN use cases, the resulting policy leads to near-optimal regret while accounting for the GNN training dynamics introduced by SGD. From a practical standpoint, this translates into lower variance estimates and competitive or superior test accuracy across several benchmarks.
引用
收藏
页数:12
相关论文
共 50 条
  • [31] Behavior and neural basis of near-optimal visual search
    Ma, Wei Ji
    Navalpakkam, Vidhya
    Beck, Jeffrey M.
    van den Berg, Ronald
    Pouget, Alexandre
    [J]. NATURE NEUROSCIENCE, 2011, 14 (06) : 783 - U150
  • [32] Behavior and neural basis of near-optimal visual search
    Wei Ji Ma
    Vidhya Navalpakkam
    Jeffrey M Beck
    Ronald van den Berg
    Alexandre Pouget
    [J]. Nature Neuroscience, 2011, 14 : 783 - 790
  • [33] Near-optimal radio use for wireless network synchronization
    Bradonjic, Milan
    Kohler, Eddie
    Ostrovsky, Rafail
    [J]. THEORETICAL COMPUTER SCIENCE, 2012, 453 : 14 - 28
  • [34] Near-Optimal Radio Use for Wireless Network Synchronization
    Bradonjic, Milan
    Kohler, Eddie
    Ostrovsky, Rafail
    [J]. ALGORITHMIC ASPECTS OF WIRELESS SENSOR NETWORKS, 2009, 5804 : 15 - +
  • [35] The Fast Loaded Dice Roller: A Near-Optimal Exact Sampler for Discrete Probability Distributions
    Saad, Feras A.
    Freer, Cameron E.
    Rinard, Martin C.
    Mansinghka, Vikash K.
    [J]. INTERNATIONAL CONFERENCE ON ARTIFICIAL INTELLIGENCE AND STATISTICS, VOL 108, 2020, 108 : 1036 - 1045
  • [36] From Near-Optimal Bayesian Integration to Neuromorphic Hardware: A Neural Network Model of Multisensory Integration
    Oess, Timo
    Loehr, Maximilian P. R.
    Schmid, Daniel
    Ernst, Marc O.
    Neumann, Heiko
    [J]. FRONTIERS IN NEUROROBOTICS, 2020, 14
  • [37] Neural Network Approximation Based Near-Optimal Motion Planning With Kinodynamic Constraints Using RRT
    Li, Yang
    Cui, Rongxin
    Li, Zhijun
    Xu, Demin
    [J]. IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS, 2018, 65 (11) : 8718 - 8729
  • [38] A fundamental study on the optimal/near-optimal shape of a network for energy distribution
    Xia, Liang
    Chan, Ming-yin
    Qu, Minglu
    Xu, Xiangguo
    Deng, Shiming
    [J]. ENERGY, 2011, 36 (11) : 6471 - 6478
  • [39] Near-Optimal and Practical Algorithms for Graph Scan Statistics with Connectivity Constraints
    Cadena, Jose
    Chen, Feng
    Vullikanti, Anil
    [J]. ACM TRANSACTIONS ON KNOWLEDGE DISCOVERY FROM DATA, 2019, 13 (02)
  • [40] Near-optimal blacklisting
    Dimitrakakis, Christos
    Mitrokotsa, Aikaterini
    [J]. COMPUTERS & SECURITY, 2017, 64 : 110 - 121