Gaussian Process Surrogate Models for Neural Networks

被引:0
|
作者
Li, Michael Y. [1 ]
Grant, Erin [2 ]
Griffiths, Thomas L. [3 ]
机构
[1] Stanford Univ, Dept Comp Sci, Stanford, CA 94305 USA
[2] UCL, Gatsby Computat Neurosci Unit, London, England
[3] Dept Psychol & Comp Sci, Princeton, NJ USA
来源
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
Not being able to understand and predict the behavior of deep learning systems makes it hard to decide what architecture and algorithm to use for a given problem. In science and engineering, modeling is a methodology used to understand complex systems whose internal processes are opaque. Modeling replaces a complex system with a simpler, more interpretable surrogate. Drawing inspiration from this, we construct a class of surrogate models for neural networks using Gaussian processes. Rather than deriving kernels for infinite neural networks, we learn kernels empirically from the naturalistic behavior of finite neural networks. We demonstrate our approach captures existing phenomena related to the spectral bias of neural networks, and then show that our surrogate models can be used to solve practical problems such as identifying which points most influence the behavior of specific neural networks and predicting which architectures and algorithms will generalize well for specific datasets.
引用
收藏
页码:1241 / 1252
页数:12
相关论文
共 50 条
  • [41] Face detection in complex background based on Gaussian models and neural networks
    Li, Qi
    Ji, Hongbing
    2006 8TH INTERNATIONAL CONFERENCE ON SIGNAL PROCESSING, VOLS 1-4, 2006, : 835 - +
  • [42] The Limitations of Large Width in Neural Networks: A Deep Gaussian Process Perspective
    Pleiss, Geoff
    Cunningham, John P.
    ADVANCES IN NEURAL INFORMATION PROCESSING SYSTEMS 34 (NEURIPS 2021), 2021,
  • [43] Probabilistic Lane Change Prediction using Gaussian Process Neural Networks
    Krueger, Martin
    Novo, Anne Stockem
    Nattermann, Till
    Bertram, Torsten
    2019 IEEE INTELLIGENT TRANSPORTATION SYSTEMS CONFERENCE (ITSC), 2019, : 3651 - 3656
  • [44] Conditional Simulation of No-Gaussian Stochastic Process Based on Neural Networks
    Li, Jinhua
    Chen, Shuisheng
    Sheng, Weibing
    ADVANCED MECHANICAL DESIGN, PTS 1-3, 2012, 479-481 : 1959 - +
  • [45] A case based comparison of identification with neural network and Gaussian process models
    Kocijan, J
    Banko, B
    Likar, B
    Girard, A
    Murray-Smith, R
    Rasmussen, CE
    INTELLIGENT CONTROL SYSTEMS AND SIGNAL PROCESSING 2003, 2003, : 129 - 134
  • [46] DESIGN OF GRADIENT NANOTWINNED METAL MATERIALS USING ADAPTIVE GAUSSIAN PROCESS BASED SURROGATE MODELS
    Zhou, Haofei
    Chen, Xin
    Li, Yumeng
    PROCEEDINGS OF THE ASME INTERNATIONAL DESIGN ENGINEERING TECHNICAL CONFERENCES AND COMPUTERS AND INFORMATION IN ENGINEERING CONFERENCE, 2019, VOL 2A, 2020,
  • [47] Using neural networks as models of personality process: A tutorial
    Read, Stephen J.
    Droutman, Vita
    Smith, Benjamin J.
    Miller, Lynn C.
    PERSONALITY AND INDIVIDUAL DIFFERENCES, 2019, 136 : 52 - 67
  • [48] NEURAL NETWORKS BASED MODELS APPLIED TO AN ELECTROCOAGULATION PROCESS
    Piuleac, Ciprian-George
    Furtuna, Renata
    Pislaru, Marius
    Curteanu, Silvia
    ENVIRONMENTAL ENGINEERING AND MANAGEMENT JOURNAL, 2011, 10 (03): : 375 - 380
  • [49] Unified estimate of Gaussian kernel width for surrogate models
    Wu, Zeping
    Wang, Donghui
    Okolo, Patrick N.
    Jiang, Zhenyu
    Zhang, Weihua
    NEUROCOMPUTING, 2016, 203 : 41 - 51
  • [50] Gaussian process as complement to test functions for surrogate modeling
    Haftka, Raphael T.
    Park, Chanyoung
    STRUCTURAL AND MULTIDISCIPLINARY OPTIMIZATION, 2020, 61 (03) : 855 - 861