Critical Initialization of Wide and Deep Neural Networks through Partial Jacobians: General Theory and Applications

• URL: http://arxiv.org/abs/2111.12143v4
• Date: Thu, 5 Oct 2023 22:44:08 GMT
• Title: Critical Initialization of Wide and Deep Neural Networks through Partial Jacobians: General Theory and Applications
• Authors: Darshil Doshi, Tianyu He, Andrey Gromov
• Abstract summary: We introduce emphpartial Jacobians of a network, defined as derivatives of preactivations in layer $l$ with respect to preactivations in layer $l_0leq l$. We derive recurrence relations for the norms of partial Jacobians and utilize these relations to analyze criticality of deep fully connected neural networks with LayerNorm and/or residual connections.
• Score: 6.579523168465526
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Deep neural networks are notorious for defying theoretical treatment. However, when the number of parameters in each layer tends to infinity, the network function is a Gaussian process (GP) and quantitatively predictive description is possible. Gaussian approximation allows one to formulate criteria for selecting hyperparameters, such as variances of weights and biases, as well as the learning rate. These criteria rely on the notion of criticality defined for deep neural networks. In this work we describe a new practical way to diagnose criticality. We introduce \emph{partial Jacobians} of a network, defined as derivatives of preactivations in layer $l$ with respect to preactivations in layer $l_0\leq l$. We derive recurrence relations for the norms of partial Jacobians and utilize these relations to analyze criticality of deep fully connected neural networks with LayerNorm and/or residual connections. We derive and implement a simple and cheap numerical test that allows one to select optimal initialization for a broad class of deep neural networks; containing fully connected, convolutional and normalization layers. Using these tools we show quantitatively that proper stacking of the LayerNorm (applied to preactivations) and residual connections leads to an architecture that is critical for any initialization. Finally, we apply our methods to analyze ResNet and MLP-Mixer architectures; demonstrating the everywhere-critical regime.

Related papers

• Wide Neural Networks as Gaussian Processes: Lessons from Deep Equilibrium Models [16.07760622196666]
  We study the deep equilibrium model (DEQ), an infinite-depth neural network with shared weight matrices across layers. Our analysis reveals that as the width of DEQ layers approaches infinity, it converges to a Gaussian process. Remarkably, this convergence holds even when the limits of depth and width are interchanged.
  arXiv  Detail & Related papers  (2023-10-16T19:00:43Z)
• Globally Optimal Training of Neural Networks with Threshold Activation Functions [63.03759813952481]
  We study weight decay regularized training problems of deep neural networks with threshold activations. We derive a simplified convex optimization formulation when the dataset can be shattered at a certain layer of the network.
  arXiv  Detail & Related papers  (2023-03-06T18:59:13Z)
• 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)
• Robust Training and Verification of Implicit Neural Networks: A Non-Euclidean Contractive Approach [64.23331120621118]
  This paper proposes a theoretical and computational framework for training and robustness verification of implicit neural networks. We introduce a related embedded network and show that the embedded network can be used to provide an $ell_infty$-norm box over-approximation of the reachable sets of the original network. We apply our algorithms to train implicit neural networks on the MNIST dataset and compare the robustness of our models with the models trained via existing approaches in the literature.
  arXiv  Detail & Related papers  (2022-08-08T03:13:24Z)
• Generalization Error Bounds for Iterative Recovery Algorithms Unfolded as Neural Networks [6.173968909465726]
  We introduce a general class of neural networks suitable for sparse reconstruction from few linear measurements. By allowing a wide range of degrees of weight-sharing between the layers, we enable a unified analysis for very different neural network types.
  arXiv  Detail & Related papers  (2021-12-08T16:17:33Z)
• Layer Adaptive Node Selection in Bayesian Neural Networks: Statistical Guarantees and Implementation Details [0.5156484100374059]
  Sparse deep neural networks have proven to be efficient for predictive model building in large-scale studies. We propose a Bayesian sparse solution using spike-and-slab Gaussian priors to allow for node selection during training. We establish the fundamental result of variational posterior consistency together with the characterization of prior parameters.
  arXiv  Detail & Related papers  (2021-08-25T00:48:07Z)
• Compressive Sensing and Neural Networks from a Statistical Learning Perspective [4.561032960211816]
  We present a generalization error analysis for a class of neural networks suitable for sparse reconstruction from few linear measurements. Under realistic conditions, the generalization error scales only logarithmically in the number of layers, and at most linear in number of measurements.
  arXiv  Detail & Related papers  (2020-10-29T15:05:43Z)
• Finite Versus Infinite Neural Networks: an Empirical Study [69.07049353209463]
  kernel methods outperform fully-connected finite-width networks. Centered and ensembled finite networks have reduced posterior variance. Weight decay and the use of a large learning rate break the correspondence between finite and infinite networks.
  arXiv  Detail & Related papers  (2020-07-31T01:57:47Z)
• Modeling from Features: a Mean-field Framework for Over-parameterized Deep Neural Networks [54.27962244835622]
  This paper proposes a new mean-field framework for over- parameterized deep neural networks (DNNs) In this framework, a DNN is represented by probability measures and functions over its features in the continuous limit. We illustrate the framework via the standard DNN and the Residual Network (Res-Net) architectures.
  arXiv  Detail & Related papers  (2020-07-03T01:37:16Z)
• MSE-Optimal Neural Network Initialization via Layer Fusion [68.72356718879428]
  Deep neural networks achieve state-of-the-art performance for a range of classification and inference tasks. The use of gradient combined nonvolutionity renders learning susceptible to novel problems. We propose fusing neighboring layers of deeper networks that are trained with random variables.
  arXiv  Detail & Related papers  (2020-01-28T18:25:15Z)

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.