With the rise of big data analytics, multi-layer neural networks have surfaced as one of the most powerful machine learning methods. However, their theoretical mathematical properties are still not fully understood. Training a neural network requires optimizing a non-convex objective function, typically done using stochastic gradient descent (SGD). In this paper, we seek to extend the mean field results of Mei et al. (2018) from two-layer neural networks with one hidden layer to three-layer neural networks with two hidden layers. We will show that the SGD dynamics is captured by a set of non-linear partial differential equations, and prove that the distributions of weights in the two hidden layers are independent. We will also detail exploratory work done based on simulation and real-world data.
EPFL（洛桑联邦理工学院）；UC San Diego（加州大学圣地亚哥分校）；UC Berkeley（加州大学伯克利分校）；Columbia University（哥伦比亚大学）
Having similar behavior at training time and test time $-$ what we call a "What You See Is What You Get" (WYSIWYG) property $-$ is desirable in machine learning. Models trained with standard stochastic gradient descent (SGD), however, do not necessarily have this property, as their complex behaviors such as robustness or subgroup performance can differ drastically between training and test time. In contrast, we show that Differentially-Private (DP) training provably ensures the high-level WYSIWYG property, which we quantify using a notion of distributional generalization. Applying this connect…
Stochastic Gradient Descent (SGD) is the most popular algorithm for training deep neural networks (DNNs). As larger networks and datasets cause longer training times, training on distributed systems is common and distributed SGD variants, mainly asynchronous and synchronous SGD, are widely used. Asynchronous SGD is communication efficient but suffers from accuracy degradation due to delayed parameter updating. Synchronous SGD becomes communication intensive when the number of nodes increases regardless of its advantage. To address these issues, we introduce Layered SGD (LSGD), a new decentrali…
University of Texas at Austin（得克萨斯大学奥斯汀分校）；Princeton University（普林斯顿大学）；Massachusetts Institute of Technology（麻省理工学院）
Generative adversarial networks (GANs) are a widely used framework for learning generative models. Wasserstein GANs (WGANs), one of the most successful variants of GANs, require solving a minmax optimization problem to global optimality, but are in practice successfully trained using stochastic gradient descent-ascent. In this paper, we show that, when the generator is a one-layer network, stochastic gradient descent-ascent converges to a global solution with polynomial time and sample complexity.
To reduce the long training time of large deep neural network (DNN) models, distributed synchronous stochastic gradient descent (S-SGD) is commonly used on a cluster of workers. However, the speedup brought by multiple workers is limited by the communication overhead. Two approaches, namely pipelining and gradient sparsification, have been separately proposed to alleviate the impact of communication overheads. Yet, the gradient sparsification methods can only initiate the communication after the backpropagation, and hence miss the pipelining opportunity. In this paper, we propose a new distrib…
University of Utah（犹他大学）；Cornell University（康奈尔大学）
The minibatch stochastic gradient descent method (SGD) is widely applied in deep learning due to its efficiency and scalability that enable training deep networks with a large volume of data. Particularly in the distributed setting, SGD is usually applied with large batch size. However, as opposed to small-batch SGD, neural network models trained with large-batch SGD can hardly generalize well, i.e., the validation accuracy is low. In this work, we introduce a novel regularization technique, namely distinctive regularization (DReg), which replicates a certain layer of the deep network and enco…
This manuscript investigates the one-pass stochastic gradient descent (SGD) dynamics of a two-layer neural network trained on Gaussian data and labels generated by a similar, though not necessarily identical, target function. We rigorously analyse the limiting dynamics via a deterministic and low-dimensional description in terms of the sufficient statistics for the population risk. Our unifying analysis bridges different regimes of interest, such as the classical gradient-flow regime of vanishing learning rate, the high-dimensional regime of large input dimension, and the overparameterised "me…
University of Minnesota, Twin Cities（明尼苏达大学双城分校）；University of Illinois Urbana-Champaign（伊利诺伊大学厄巴纳-香槟分校）
In spite of advances in understanding lazy training, recent work attributes the practical success of deep learning to the rich regime with complex inductive bias. In this paper, we study rich regime training empirically with benchmark datasets, and find that while most parameters are lazy, there is always a small number of active parameters which change quite a bit during training. We show that re-initializing (resetting to their initial random values) the active parameters leads to worse generalization. Further, we show that most of the active parameters are in the bottom layers, close to the…
Stochastic gradient descent (SGD) has achieved great success in training deep neural network, where the gradient is computed through back-propagation. However, the back-propagated values of different layers vary dramatically. This inconsistence of gradient magnitude across different layers renders optimization of deep neural network with a single learning rate problematic. We introduce the back-matching propagation which computes the backward values on the layer's parameter and the input by matching backward values on the layer's output. This leads to solving a bunch of least-squares problems,…
Computer Science, Wayne State University（韦恩州立大学计算机科学（系））
Deep neural networks (DNNs) for supervised learning can be viewed as a pipeline of a feature extractor (i.e. last hidden layer) and a linear classifier (i.e. output layer) that is trained jointly with stochastic gradient descent (SGD). In each iteration of SGD, a mini-batch from the training data is sampled and the true gradient of the loss function is estimated as the noisy gradient calculated on this mini-batch. From the feature learning perspective, the feature extractor should be updated to learn meaningful features with respect to the entire data, and reduce the accommodation to noise in …
Artificial intelligence is revolutionizing our lives at an ever increasing pace. At the heart of this revolution is the recent advancements in deep neural networks (DNN), learning to perform sophisticated, high-level tasks. However, training DNNs requires massive amounts of data and is very computationally intensive. Gaining analytical understanding of the solutions found by DNNs can help us devise more efficient training algorithms, replacing the commonly used mthod of stochastic gradient descent (SGD). We analyze the dynamics of SGD and show that, indeed, direct computation of the solutions …
Multi-layer neural networks are among the most powerful models in machine learning, yet the fundamental reasons for this success defy mathematical understanding. Learning a neural network requires to optimize a non-convex high-dimensional objective (risk function), a problem which is usually attacked using stochastic gradient descent (SGD). Does SGD converge to a global optimum of the risk or only to a local optimum? In the first case, does this happen because local minima are absent, or because SGD somehow avoids them? In the second, why do local minima reached by SGD have good generalization…
Université de Montréal（蒙特利尔大学）；Mila Institute（米拉研究所）；University of Toronto（多伦多大学）；Vector Institute（矢量研究所）；Korea University（高丽大学）；Saint Mary’s University（圣玛丽大学）
We study the problem of training a two-layer neural network (NN) of arbitrary width using stochastic gradient descent (SGD) where the input $\boldsymbol{x}\in \mathbb{R}^d$ is Gaussian and the target $y \in \mathbb{R}$ follows a multiple-index model, i.e., $y=g(\langle\boldsymbol{u_1},\boldsymbol{x}\rangle,...,\langle\boldsymbol{u_k},\boldsymbol{x}\rangle)$ with a noisy link function $g$. We prove that the first-layer weights of the NN converge to the $k$-dimensional principal subspace spanned by the vectors $\boldsymbol{u_1},...,\boldsymbol{u_k}$ of the true model, when online SGD with weight…
In recent years, stochastic gradient descent (SGD) based techniques has become the standard tools for training neural networks. However, formal theoretical understanding of why SGD can train neural networks in practice is largely missing. In this paper, we make progress on understanding this mystery by providing a convergence analysis for SGD on a rich subset of two-layer feedforward networks with ReLU activations. This subset is characterized by a special structure called "identity mapping". We prove that, if input follows from Gaussian distribution, with standard $O(1/\sqrt{d})$ initializati…
Stochastic gradient descent (SGD) has been the dominant optimization method for training deep neural networks due to its many desirable properties. One of the more remarkable and least understood quality of SGD is that it generalizes relatively well on unseen data even when the neural network has millions of parameters. We hypothesize that in certain cases it is desirable to relax its intrinsic generalization properties and introduce an extension of SGD called deep gradient boosting (DGB). The key idea of DGB is that back-propagated gradients inferred using the chain rule can be viewed as pseu…
The redundancy is widely recognized in Convolutional Neural Networks (CNNs), which enables to remove unimportant filters from convolutional layers so as to slim the network with acceptable performance drop. Inspired by the linear and combinational properties of convolution, we seek to make some filters increasingly close and eventually identical for network slimming. To this end, we propose Centripetal SGD (C-SGD), a novel optimization method, which can train several filters to collapse into a single point in the parameter hyperspace. When the training is completed, the removal of the identica…
Second-order optimization methods such as natural gradient descent have the potential to speed up training of neural networks by correcting for the curvature of the loss function. Unfortunately, the exact natural gradient is impractical to compute for large models, and most approximations either require an expensive iterative procedure or make crude approximations to the curvature. We present Kronecker Factors for Convolution (KFC), a tractable approximation to the Fisher matrix for convolutional networks based on a structured probabilistic model for the distribution over backpropagated deriva…
University of New Brunswick（新不伦瑞克大学）；University of Toronto（多伦多大学）
Explaining the generalization characteristics of deep learning is an emerging topic in advanced machine learning. There are several unanswered questions about how learning under stochastic optimization really works and why certain strategies are better than others. In this paper, we address the following question: \textit{can we probe intermediate layers of a deep neural network to identify and quantify the learning quality of each layer?} With this question in mind, we propose new explainability metrics that measure the redundant information in a network's layers using a low-rank factorizatio…
Deep unfolding networks have recently gained popularity in the context of solving imaging inverse problems. However, the computational and memory complexity of data-consistency layers within traditional deep unfolding networks scales with the number of measurements, limiting their applicability to large-scale imaging inverse problems. We propose SGD-Net as a new methodology for improving the efficiency of deep unfolding through stochastic approximations of the data-consistency layers. Our theoretical analysis shows that SGD-Net can be trained to approximate batch deep unfolding networks to an …
The merged-staircase property: a necessary and nearly sufficient condition for SGD learning of sparse functions on two-layer neural networks
Mathematics Institute, EPFL（EPFL数学研究所）；Department of Electrical Engineering and Computer Science, MIT（MIT电气工程与计算机科学系）；Department of Statistics, Stanford University（斯坦福大学统计系）
University of California, Los Angeles（加利福尼亚大学洛杉矶分校）
We consider a one-hidden-layer leaky ReLU network of arbitrary width trained by stochastic gradient descent (SGD) following an arbitrary initialization. We prove that SGD produces neural networks that have classification accuracy competitive with that of the best halfspace over the distribution for a broad class of distributions that includes log-concave isotropic and hard margin distributions. Equivalently, such networks can generalize when the data distribution is linearly separable but corrupted with adversarial label noise, despite the capacity to overfit. To the best of our knowledge, thi…
On the Learning Dynamics of Two-layer Linear Networks with Label Noise SGD
Sch. of Computer Science and Zhiyuan College, Shanghai Jiao Tong University（上海交通大学计算机科学学院和智远学院）；Peking University（北京大学）；University of Sydney（悉尼大学）；RIKEN Center for Advanced Intelligence Project（理化学研究所先进智能项目中心）；The Institute of Statistical Mathematics（统计数学研究所）；The University of Tokyo（东京大学）
Institute of Science and Technology, Austria（奥地利科学技术研究所）
The optimization of multilayer neural networks typically leads to a solution with zero training error, yet the landscape can exhibit spurious local minima and the minima can be disconnected. In this paper, we shed light on this phenomenon: we show that the combination of stochastic gradient descent (SGD) and over-parameterization makes the landscape of multilayer neural networks approximately connected and thus more favorable to optimization. More specifically, we prove that SGD solutions are connected via a piecewise linear path, and the increase in loss along this path vanishes as the number…
Centre Borelli（博雷利中心）；ENS Paris Saclay（巴黎-萨克雷高等师范学校）；Télécom Paris（巴黎电信学院）；Institut Polytechnique de Paris（巴黎综合理工学院）
In this paper, we investigate the limiting behavior of a continuous-time counterpart of the Stochastic Gradient Descent (SGD) algorithm applied to two-layer overparameterized neural networks, as the number or neurons (ie, the size of the hidden layer) $N \to +\infty$. Following a probabilistic approach, we show 'propagation of chaos' for the particle system defined by this continuous-time dynamics under different scenarios, indicating that the statistical interaction between the particles asymptotically vanishes. In particular, we establish quantitative convergence with respect to $N$ of any p…
The multiplicative structure of parameters and input data in the first layer of neural networks is explored to build connection between the landscape of the loss function with respect to parameters and the landscape of the model function with respect to input data. By this connection, it is shown that flat minima regularize the gradient of the model function, which explains the good generalization performance of flat minima. Then, we go beyond the flatness and consider high-order moments of the gradient noise, and show that Stochastic Gradient Descent (SGD) tends to impose constraints on these…
University of California, Davis（加利福尼亚大学戴维斯分校）；University of California, Los Angles（加利福尼亚大学洛杉矶分校）
In deep neural nets, lower level embedding layers account for a large portion of the total number of parameters. Tikhonov regularization, graph-based regularization, and hard parameter sharing are approaches that introduce explicit biases into training in a hope to reduce statistical complexity. Alternatively, we propose stochastically shared embeddings (SSE), a data-driven approach to regularizing embedding layers, which stochastically transitions between embeddings during stochastic gradient descent (SGD). Because SSE integrates seamlessly with existing SGD algorithms, it can be used with on…
SGD Provably Prioritizes a Shortcut Spurious Feature in the XOR Model
H. Milton Stewart School of Industrial and Systems Engineering, Georgia Institute of Technology（霍华德·米尔顿·斯图尔特工业与系统工程学院，佐治亚理工学院）；School of Electrical and Computer Engineering, Georgia Institute of Technology（电气与计算机工程学院，佐治亚理工学院）
Optimization Risk Bounds for Kolmogorov-Arnold Networks Trained by DP-SGD with Correlated Noise
RPTU Kaiserslautern-Landau（凯撒斯劳滕-兰道大学）；Machine Learning Research, Morgan Stanley（摩根大通机器学习研究部）；Institute of Science and Technology, Klosterneuburg（克洛斯特纽堡科学与技术研究所）；Catholic University of Eichstätt-Ingolstadt（埃希施泰特-英戈尔施塔特天主教大学）
Google DeepMind Gatsby Unit, University College London（谷歌深Mind加布兹单位，伦敦大学学院）；Tübingen AI Center, University of Tübingen（图宾根人工智能中心，图宾根大学）；University College London（伦敦大学学院）
On the Superlinear Relationship between SGD Noise Covariance and Loss Landscape Curvature
School of Physics, Peking University（北京大学物理系）；Peking University Chengdu Academy for Advanced Interdisciplinary Biotechnologies（北京大学成都先进交叉生物技术研究院）；Flatiron Institute（Flatiron研究所）
University of Münster（明斯特大学）；The Chinese University of Hong Kong, Shenzhen（香港中文大学（深圳））
In this article we study the stochastic gradient descent (SGD) optimization method in the training of fully-connected feedforward artificial neural networks with ReLU activation. The main result of this work proves that the risk of the SGD process converges to zero if the target function under consideration is constant. In the established convergence result the considered artificial neural networks consist of one input layer, one hidden layer, and one output layer (with $d \in \mathbb{N}$ neurons on the input layer, $H \in \mathbb{N}$ neurons on the hidden layer, and one neuron on the output l…
University of California, Los Angeles（加利福尼亚大学洛杉矶分校）；Google（谷歌）
We study the convergence of gradient descent (GD) and stochastic gradient descent (SGD) for training $L$-hidden-layer linear residual networks (ResNets). We prove that for training deep residual networks with certain linear transformations at input and output layers, which are fixed throughout training, both GD and SGD with zero initialization on all hidden weights can converge to the global minimum of the training loss. Moreover, when specializing to appropriate Gaussian random linear transformations, GD and SGD provably optimize wide enough deep linear ResNets. Compared with the global conve…
There has been a recent surge of interest in understanding the convergence of gradient descent (GD) and stochastic gradient descent (SGD) in overparameterized neural networks. Most previous works assume that the training data is provided a priori in a batch, while less attention has been paid to the important setting where the training data arrives in a stream. In this paper, we study the streaming data setup and show that with overparamterization and random initialization, the prediction error of two-layer neural networks under one-pass SGD converges in expectation. The convergence rate depen…
This paper follows up on a recent work of Neu et al. (2021) and presents some new information-theoretic upper bounds for the generalization error of machine learning models, such as neural networks, trained with SGD. We apply these bounds to analyzing the generalization behaviour of linear and two-layer ReLU networks. Experimental study of these bounds provide some insights on the SGD training of neural networks. They also point to a new and simple regularization scheme which we show performs comparably to the current state of the art.
Spectral alignment of stochastic gradient descent for high-dimensional classification tasks
Courant Institute, New York University（纽约大学Courant研究所）；Department of Mathematics, Northwestern University（西北大学数学系）；Department of Statistics and Data Science, University of Pennsylvania（宾夕法尼亚大学统计与数据科学系）；Department of Statistics and Actuarial Science, Department of Applied Mathematics, and Cheriton School of Computer Science, University of Waterloo（滑铁卢大学统计与精算系、应用数学系及Cheriton计算机科学学院）
Distributing Neural Network training is of particular interest for several reasons including scaling using computing clusters, training at data sources such as IOT devices and edge servers, utilizing underutilized resources across heterogeneous environments, and so on. Most contemporary approaches primarily address scaling using computing clusters and require high network bandwidth and frequent communication. This thesis presents an overview of standard approaches to distribute training and proposes a novel technique involving pairwise-communication using Gossip-like protocols, called Elastic …
The Blavatnik School of Computer Science, Tel Aviv University（特拉维夫大学布拉瓦特尼克计算机科学学院）；School of Computer Science and Engineering, The Hebrew University（希伯来大学计算机科学与工程学院）
Neural networks exhibit good generalization behavior in the over-parameterized regime, where the number of network parameters exceeds the number of observations. Nonetheless, current generalization bounds for neural networks fail to explain this phenomenon. In an attempt to bridge this gap, we study the problem of learning a two-layer over-parameterized neural network, when the data is generated by a linearly separable function. In the case where the network has Leaky ReLU activations, we provide both optimization and generalization guarantees for over-parameterized networks. Specifically, we …
Neural networks trained with SGD learn distributions of increasing complexity
Université PSL（巴黎文理研究大学）；CNRS（法国国家科学研究中心）；Sorbonne Université（索邦大学）；Université Paris-Diderot（巴黎第七大学）；École Fédérale Polytechnique de Lausanne (EPFL)（洛桑联邦理工学院）；The Abdus Salam International Centre for Theoretical Physics (ICTP)（阿卜杜勒·萨拉姆国际理论物理中心）；International School of Advanced Studies (SISSA)（国际高等研究院）
This work establishes low test error of gradient flow (GF) and stochastic gradient descent (SGD) on two-layer ReLU networks with standard initialization, in three regimes where key sets of weights rotate little (either naturally due to GF and SGD, or due to an artificial constraint), and making use of margins as the core analytic technique. The first regime is near initialization, specifically until the weights have moved by $\mathcal{O}(\sqrt m)$, where $m$ denotes the network width, which is in sharp contrast to the $\mathcal{O}(1)$ weight motion allowed by the Neural Tangent Kernel (NTK); h…
