Learning from Randomly Initialized Neural Network Features

• URL: http://arxiv.org/abs/2202.06438v1
• Date: Sun, 13 Feb 2022 23:35:11 GMT
• Title: Learning from Randomly Initialized Neural Network Features
• Authors: Ehsan Amid, Rohan Anil, Wojciech Kot{\l}owski, Manfred K. Warmuth
• Abstract summary: We present the surprising result that randomly neural networks are good feature extractors in expectation. These random features correspond to finite-sample realizations of what we call Neural Network Prior Kernel (NNPK), which is inherently infinite-dimensional.
• Score: 24.75062551820944
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: We present the surprising result that randomly initialized neural networks are good feature extractors in expectation. These random features correspond to finite-sample realizations of what we call Neural Network Prior Kernel (NNPK), which is inherently infinite-dimensional. We conduct ablations across multiple architectures of varying sizes as well as initializations and activation functions. Our analysis suggests that certain structures that manifest in a trained model are already present at initialization. Therefore, NNPK may provide further insight into why neural networks are so effective in learning such structures.

Related papers

• Convergence Analysis for Deep Sparse Coding via Convolutional Neural Networks [7.956678963695681]
  We introduce a novel class of Deep Sparse Coding (DSC) models. We derive convergence rates for CNNs in their ability to extract sparse features. Inspired by the strong connection between sparse coding and CNNs, we explore training strategies to encourage neural networks to learn more sparse features.
  arXiv  Detail & Related papers  (2024-08-10T12:43:55Z)
• How neural networks learn to classify chaotic time series [77.34726150561087]
  We study the inner workings of neural networks trained to classify regular-versus-chaotic time series. We find that the relation between input periodicity and activation periodicity is key for the performance of LKCNN models.
  arXiv  Detail & Related papers  (2023-06-04T08:53:27Z)
• Gradient Descent in Neural Networks as Sequential Learning in RKBS [63.011641517977644]
  We construct an exact power-series representation of the neural network in a finite neighborhood of the initial weights. We prove that, regardless of width, the training sequence produced by gradient descent can be exactly replicated by regularized sequential learning.
  arXiv  Detail & Related papers  (2023-02-01T03:18:07Z)
• Neural Networks with Sparse Activation Induced by Large Bias: Tighter Analysis with Bias-Generalized NTK [86.45209429863858]
  We study training one-hidden-layer ReLU networks in the neural tangent kernel (NTK) regime. We show that the neural networks possess a different limiting kernel which we call textitbias-generalized NTK We also study various properties of the neural networks with this new kernel.
  arXiv  Detail & Related papers  (2023-01-01T02:11:39Z)
• Dynamical systems' based neural networks [0.7874708385247353]
  We build neural networks using a suitable, structure-preserving, numerical time-discretisation. The structure of the neural network is then inferred from the properties of the ODE vector field. We present two universal approximation results and demonstrate how to impose some particular properties on the neural networks.
  arXiv  Detail & Related papers  (2022-10-05T16:30:35Z)
• Extrapolation and Spectral Bias of Neural Nets with Hadamard Product: a Polynomial Net Study [55.12108376616355]
  The study on NTK has been devoted to typical neural network architectures, but is incomplete for neural networks with Hadamard products (NNs-Hp) In this work, we derive the finite-width-K formulation for a special class of NNs-Hp, i.e., neural networks. We prove their equivalence to the kernel regression predictor with the associated NTK, which expands the application scope of NTK.
  arXiv  Detail & Related papers  (2022-09-16T06:36:06Z)
• On the Neural Tangent Kernel Analysis of Randomly Pruned Neural Networks [91.3755431537592]
  We study how random pruning of the weights affects a neural network's neural kernel (NTK) In particular, this work establishes an equivalence of the NTKs between a fully-connected neural network and its randomly pruned version.
  arXiv  Detail & Related papers  (2022-03-27T15:22:19Z)
• Neural networks with linear threshold activations: structure and algorithms [1.795561427808824]
  We show that 2 hidden layers are necessary and sufficient to represent any function representable in the class. We also give precise bounds on the sizes of the neural networks required to represent any function in the class. We propose a new class of neural networks that we call shortcut linear threshold networks.
  arXiv  Detail & Related papers  (2021-11-15T22:33:52Z)
• What can linearized neural networks actually say about generalization? [67.83999394554621]
  In certain infinitely-wide neural networks, the neural tangent kernel (NTK) theory fully characterizes generalization. We show that the linear approximations can indeed rank the learning complexity of certain tasks for neural networks. Our work provides concrete examples of novel deep learning phenomena which can inspire future theoretical research.
  arXiv  Detail & Related papers  (2021-06-12T13:05:11Z)
• Expressivity of Deep Neural Networks [2.7909470193274593]
  In this review paper, we give a comprehensive overview of the large variety of approximation results for neural networks. While the mainbody of existing results is for general feedforward architectures, we also depict approximation results for convolutional, residual and recurrent neural networks.
  arXiv  Detail & Related papers  (2020-07-09T13:08:01Z)
• Deep Randomized Neural Networks [12.333836441649343]
  Randomized Neural Networks explore the behavior of neural systems where the majority of connections are fixed. This chapter surveys all the major aspects regarding the design and analysis of Randomized Neural Networks.
  arXiv  Detail & Related papers  (2020-02-27T17:57:58Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.