arXiv subjects

Michael W. Mahoney

Publications and source records attributed to Michael W. Mahoney.

At least 217 records · Page 12Linked to original sources

arXiv2019

Bootstrapping the Operator Norm in High Dimensions: Error Estimation for Covariance Matrices and Sketching

Although the operator (spectral) norm is one of the most widely used metrics for covariance estimation, comparatively little is known about the fluctuations of error in this norm. To be specific, let $\hatΣ$ denote the sample covariance matrix of $n$ observations in $\mathbb{R}^p$ that arise from a population matrix $Σ$, and let $T_n=\sqrt{n}\|\hatΣ-Σ\|_{\text{op}}$. In the setting where the eigenvalues of $Σ$ have a decay profile of the form $λ_j(Σ)\asymp j^{-2β}$, we analyze how well the bootstrap can approximate the distribution of $T_n$. Our main result shows that up to factors of $\log(n)$, the bootstrap can approximate the distribution of $T_n$ at the dimension-free rate of $n^{-\frac{β-1/2}{6β+4}}$, with respect to the Kolmogorov metric. Perhaps surprisingly, a result of this type appears to be new even in settings where $p< n$. More generally, we discuss the consequences of this result beyond covariance matrices and show how the bootstrap can be used to estimate the errors of sketching algorithms in randomized numerical linear algebra (RandNLA). An illustration of these ideas is also provided with a climate data example.

math.ST↗

arXiv2019

The Difficulties of Addressing Interdisciplinary Challenges at the Foundations of Data Science

The National Science Foundation's Transdisciplinary Research in Principles of Data Science (TRIPODS) program aims to integrate three areas central to the foundations of data by uniting the statistics, mathematics, and theoretical computer science research communities. The program aims to provide a model for funding cross-cutting research and facilitating interactions among the three disciplines. Challenges associated with orchestrating fruitful interactions are described.

cs.CY↗

arXiv2019

Inefficiency of K-FAC for Large Batch Size Training

In stochastic optimization, using large batch sizes during training can leverage parallel resources to produce faster wall-clock training times per training epoch. However, for both training loss and testing error, recent results analyzing large batch Stochastic Gradient Descent (SGD) have found sharp diminishing returns, beyond a certain critical batch size. In the hopes of addressing this, it has been suggested that the Kronecker-Factored Approximate Curvature (\mbox{K-FAC}) method allows for greater scalability to large batch sizes, for non-convex machine learning problems such as neural network optimization, as well as greater robustness to variation in model hyperparameters. Here, we perform a detailed empirical analysis of large batch size training %of these two hypotheses, for both \mbox{K-FAC} and SGD, evaluating performance in terms of both wall-clock time and aggregate computational cost. Our main results are twofold: first, we find that both \mbox{K-FAC} and SGD doesn't have ideal scalability behavior beyond a certain batch size, and that \mbox{K-FAC} does not exhibit improved large-batch scalability behavior, as compared to SGD; and second, we find that \mbox{K-FAC}, in addition to requiring more hyperparameters to tune, suffers from similar hyperparameter sensitivity behavior as does SGD. We discuss extensive results using ResNet and AlexNet on \mbox{CIFAR-10} and SVHN, respectively, as well as more general implications of our findings.

cs.LG↗

arXiv2019

Bayesian experimental design using regularized determinantal point processes

In experimental design, we are given $n$ vectors in $d$ dimensions, and our goal is to select $k\ll n$ of them to perform expensive measurements, e.g., to obtain labels/responses, for a linear regression task. Many statistical criteria have been proposed for choosing the optimal design, with popular choices including A- and D-optimality. If prior knowledge is given, typically in the form of a $d\times d$ precision matrix $\mathbf A$, then all of the criteria can be extended to incorporate that information via a Bayesian framework. In this paper, we demonstrate a new fundamental connection between Bayesian experimental design and determinantal point processes, the latter being widely used for sampling diverse subsets of data. We use this connection to develop new efficient algorithms for finding $(1+ε)$-approximations of optimal designs under four optimality criteria: A, C, D and V. Our algorithms can achieve this when the desired subset size $k$ is $Ω(\frac{d_{\mathbf A}}ε + \frac{\log 1/ε}{ε^2})$, where $d_{\mathbf A}\leq d$ is the $\mathbf A$-effective dimension, which can often be much smaller than $d$. Our results offer direct improvements over a number of prior works, for both Bayesian and classical experimental design, in terms of algorithm efficiency, approximation quality, and range of applicable criteria.

cs.LG↗

arXiv2019

Parallel Local Graph Clustering

Graph clustering has many important applications in computing, but due to growing sizes of graphs, even traditionally fast clustering methods such as spectral partitioning can be computationally expensive for real-world graphs of interest. Motivated partly by this, so-called local algorithms for graph clustering have received significant interest due to the fact that they can find good clusters in a graph with work proportional to the size of the cluster rather than that of the entire graph. This feature has proven to be crucial in making such graph clustering and many of its downstream applications efficient in practice. While local clustering algorithms are already faster than traditional algorithms that touch the entire graph, they are sequential and there is an opportunity to make them even more efficient via parallelization. In this paper, we show how to parallelize many of these algorithms in the shared-memory multicore setting, and we analyze the parallel complexity of these algorithms. We present comprehensive experiments on large-scale graphs showing that our parallel algorithms achieve good parallel speedups on a modern multicore machine, thus significantly speeding up the analysis of local graph clusters in the very large-scale setting.

cs.DC↗

arXiv2019

Residual Networks as Nonlinear Systems: Stability Analysis using Linearization

We regard pre-trained residual networks (ResNets) as nonlinear systems and use linearization, a common method used in the qualitative analysis of nonlinear systems, to understand the behavior of the networks under small perturbations of the input images. We work with ResNet-56 and ResNet-110 trained on the CIFAR-10 data set. We linearize these networks at the level of residual units and network stages, and the singular value decomposition is used in the stability analysis of these components. It is found that most of the singular values of the linearizations of residual units are 1 and, in spite of the fact that the linearizations depend directly on the activation maps, the singular values differ only slightly for different input images. However, adjusting the scaling of the skip connection or the values of the weights in a residual unit has a significant impact on the singular value distributions. Inspection of how random and adversarial perturbations of input images propagate through the network reveals that there is a dramatic jump in the magnitude of adversarial perturbations towards the end of the final stage of the network that is not present in the case of random perturbations. We attempt to gain a better understanding of this phenomenon by projecting the perturbations onto singular vectors of the linearizations of the residual units.

cs.LG↗

arXiv2019

Distributed estimation of the inverse Hessian by determinantal averaging

In distributed optimization and distributed numerical linear algebra, we often encounter an inversion bias: if we want to compute a quantity that depends on the inverse of a sum of distributed matrices, then the sum of the inverses does not equal the inverse of the sum. An example of this occurs in distributed Newton's method, where we wish to compute (or implicitly work with) the inverse Hessian multiplied by the gradient. In this case, locally computed estimates are biased, and so taking a uniform average will not recover the correct solution. To address this, we propose determinantal averaging, a new approach for correcting the inversion bias. This approach involves reweighting the local estimates of the Newton's step proportionally to the determinant of the local Hessian estimate, and then averaging them together to obtain an improved global estimate. This method provides the first known distributed Newton step that is asymptotically consistent, i.e., it recovers the exact step in the limit as the number of distributed partitions grows to infinity. To show this, we develop new expectation identities and moment bounds for the determinant and adjugate of a random matrix. Determinantal averaging can be applied not only to Newton's method, but to computing any quantity that is a linear tranformation of a matrix inverse, e.g., taking a trace of the inverse covariance matrix, which is used in data uncertainty quantification.

cs.LG↗

arXiv2019

Physics-informed Autoencoders for Lyapunov-stable Fluid Flow Prediction

In addition to providing high-profile successes in computer vision and natural language processing, neural networks also provide an emerging set of techniques for scientific problems. Such data-driven models, however, typically ignore physical insights from the scientific system under consideration. Among other things, a physics-informed model formulation should encode some degree of stability or robustness or well-conditioning (in that a small change of the input will not lead to drastic changes in the output), characteristic of the underlying scientific problem. We investigate whether it is possible to include physics-informed prior knowledge for improving the model quality (e.g., generalization performance, sensitivity to parameter tuning, or robustness in the presence of noisy data). To that extent, we focus on the stability of an equilibrium, one of the most basic properties a dynamic system can have, via the lens of Lyapunov analysis. For the prototypical problem of fluid flow prediction, we show that models preserving Lyapunov stability improve the generalization error and reduce the prediction uncertainty.

physics.comp-ph↗

arXiv2019

Newton-Type Methods for Non-Convex Optimization Under Inexact Hessian Information

We consider variants of trust-region and cubic regularization methods for non-convex optimization, in which the Hessian matrix is approximated. Under mild conditions on the inexact Hessian, and using approximate solution of the corresponding sub-problems, we provide iteration complexity to achieve $ ε$-approximate second-order optimality which have shown to be tight. Our Hessian approximation conditions constitute a major relaxation over the existing ones in the literature. Consequently, we are able to show that such mild conditions allow for the construction of the approximate Hessian through various random sampling methods. In this light, we consider the canonical problem of finite-sum minimization, provide appropriate uniform and non-uniform sub-sampling strategies to construct such Hessian approximations, and obtain optimal iteration complexity for the corresponding sub-sampled trust-region and cubic regularization methods.

math.OC↗

arXiv2019

JumpReLU: A Retrofit Defense Strategy for Adversarial Attacks

It has been demonstrated that very simple attacks can fool highly-sophisticated neural network architectures. In particular, so-called adversarial examples, constructed from perturbations of input data that are small or imperceptible to humans but lead to different predictions, may lead to an enormous risk in certain critical applications. In light of this, there has been a great deal of work on developing adversarial training strategies to improve model robustness. These training strategies are very expensive, in both human and computational time. To complement these approaches, we propose a very simple and inexpensive strategy which can be used to ``retrofit'' a previously-trained network to improve its resilience to adversarial attacks. More concretely, we propose a new activation function---the JumpReLU---which, when used in place of a ReLU in an already-trained model, leads to a trade-off between predictive accuracy and robustness. This trade-off is controlled by the jump size, a hyper-parameter which can be tuned during the validation stage. Our empirical results demonstrate that this increases model robustness, protecting against adversarial attacks with substantially increased levels of perturbations. This is accomplished simply by retrofitting existing networks with our JumpReLU activation function, without the need for retraining the model. Additionally, we demonstrate that adversarially trained (robust) models can greatly benefit from retrofitting.

cs.CR↗

arXiv2019

A Bootstrap Method for Error Estimation in Randomized Matrix Multiplication

In recent years, randomized methods for numerical linear algebra have received growing interest as a general approach to large-scale problems. Typically, the essential ingredient of these methods is some form of randomized dimension reduction, which accelerates computations, but also creates random approximation error. In this way, the dimension reduction step encodes a tradeoff between cost and accuracy. However, the exact numerical relationship between cost and accuracy is typically unknown, and consequently, it may be difficult for the user to precisely know (1) how accurate a given solution is, or (2) how much computation is needed to achieve a given level of accuracy. In the current paper, we study randomized matrix multiplication (sketching) as a prototype setting for addressing these general problems. As a solution, we develop a bootstrap method for \emph{directly estimating} the accuracy as a function of the reduced dimension (as opposed to deriving worst-case bounds on the accuracy in terms of the reduced dimension). From a computational standpoint, the proposed method does not substantially increase the cost of standard sketching methods, and this is made possible by an "extrapolation" technique. In addition, we provide both theoretical and empirical results to demonstrate the effectiveness of the proposed method.

stat.ML↗

arXiv2019

Rethinking generalization requires revisiting old ideas: statistical mechanics approaches and complex learning behavior

We describe an approach to understand the peculiar and counterintuitive generalization properties of deep neural networks. The approach involves going beyond worst-case theoretical capacity control frameworks that have been popular in machine learning in recent years to revisit old ideas in the statistical mechanics of neural networks. Within this approach, we present a prototypical Very Simple Deep Learning (VSDL) model, whose behavior is controlled by two control parameters, one describing an effective amount of data, or load, on the network (that decreases when noise is added to the input), and one with an effective temperature interpretation (that increases when algorithms are early stopped). Using this model, we describe how a very simple application of ideas from the statistical mechanics theory of generalization provides a strong qualitative description of recently-observed empirical results regarding the inability of deep neural networks not to overfit training data, discontinuous learning and sharp transitions in the generalization properties of learning algorithms, etc.

cs.LG↗

arXiv2019

Scalable Kernel K-Means Clustering with Nystrom Approximation: Relative-Error Bounds

Kernel $k$-means clustering can correctly identify and extract a far more varied collection of cluster structures than the linear $k$-means clustering algorithm. However, kernel $k$-means clustering is computationally expensive when the non-linear feature map is high-dimensional and there are many input points. Kernel approximation, e.g., the Nyström method, has been applied in previous works to approximately solve kernel learning problems when both of the above conditions are present. This work analyzes the application of this paradigm to kernel $k$-means clustering, and shows that applying the linear $k$-means clustering algorithm to $\frac{k}ε (1 + o(1))$ features constructed using a so-called rank-restricted Nyström approximation results in cluster assignments that satisfy a $1 + ε$ approximation ratio in terms of the kernel $k$-means cost function, relative to the guarantee provided by the same algorithm without the use of the Nyström method. As part of the analysis, this work establishes a novel $1 + ε$ relative-error trace norm guarantee for low-rank approximation using the rank-restricted Nyström approximation. Empirical evaluations on the $8.1$ million instance MNIST8M dataset demonstrate the scalability and usefulness of kernel $k$-means clustering with Nyström approximation. This work argues that spectral clustering using Nyström approximation---a popular and computationally efficient, but theoretically unsound approach to non-linear clustering---should be replaced with the efficient and theoretically sound combination of kernel $k$-means clustering with Nyström approximation. The superior performance of the latter approach is empirically verified.

cs.LG↗

arXiv2019

Minimax experimental design: Bridging the gap between statistical and worst-case approaches to least squares regression

In experimental design, we are given a large collection of vectors, each with a hidden response value that we assume derives from an underlying linear model, and we wish to pick a small subset of the vectors such that querying the corresponding responses will lead to a good estimator of the model. A classical approach in statistics is to assume the responses are linear, plus zero-mean i.i.d. Gaussian noise, in which case the goal is to provide an unbiased estimator with smallest mean squared error (A-optimal design). A related approach, more common in computer science, is to assume the responses are arbitrary but fixed, in which case the goal is to estimate the least squares solution using few responses, as quickly as possible, for worst-case inputs. Despite many attempts, characterizing the relationship between these two approaches has proven elusive. We address this by proposing a framework for experimental design where the responses are produced by an arbitrary unknown distribution. We show that there is an efficient randomized experimental design procedure that achieves strong variance bounds for an unbiased estimator using few responses in this general model. Nearly tight bounds for the classical A-optimality criterion, as well as improved bounds for worst-case responses, emerge as special cases of this result. In the process, we develop a new algorithm for a joint sampling distribution called volume sampling, and we propose a new i.i.d. importance sampling method: inverse score sampling. A key novelty of our analysis is in developing new expected error bounds for worst-case regression by controlling the tail behavior of i.i.d. sampling via the jointness of volume sampling. Our result motivates a new minimax-optimality criterion for experimental design which can be viewed as an extension of both A-optimal design and sampling for worst-case regression.

cs.LG↗

arXiv2019

Traditional and Heavy-Tailed Self Regularization in Neural Network Models

Random Matrix Theory (RMT) is applied to analyze the weight matrices of Deep Neural Networks (DNNs), including both production quality, pre-trained models such as AlexNet and Inception, and smaller models trained from scratch, such as LeNet5 and a miniature-AlexNet. Empirical and theoretical results clearly indicate that the empirical spectral density (ESD) of DNN layer matrices displays signatures of traditionally-regularized statistical models, even in the absence of exogenously specifying traditional forms of regularization, such as Dropout or Weight Norm constraints. Building on recent results in RMT, most notably its extension to Universality classes of Heavy-Tailed matrices, we develop a theory to identify \emph{5+1 Phases of Training}, corresponding to increasing amounts of \emph{Implicit Self-Regularization}. For smaller and/or older DNNs, this Implicit Self-Regularization is like traditional Tikhonov regularization, in that there is a `size scale' separating signal from noise. For state-of-the-art DNNs, however, we identify a novel form of \emph{Heavy-Tailed Self-Regularization}, similar to the self-organization seen in the statistical physics of disordered systems. This implicit Self-Regularization can depend strongly on the many knobs of the training process. By exploiting the generalization gap phenomena, we demonstrate that we can cause a small model to exhibit all 5+1 phases of training simply by changing the batch size.

cs.LG↗

arXiv2018

Hessian-based Analysis of Large Batch Training and Robustness to Adversaries

Large batch size training of Neural Networks has been shown to incur accuracy loss when trained with the current methods. The exact underlying reasons for this are still not completely understood. Here, we study large batch size training through the lens of the Hessian operator and robust optimization. In particular, we perform a Hessian based study to analyze exactly how the landscape of the loss function changes when training with large batch size. We compute the true Hessian spectrum, without approximation, by back-propagating the second derivative. Extensive experiments on multiple networks show that saddle-points are not the cause for generalization gap of large batch size training, and the results consistently show that large batch converges to points with noticeably higher Hessian spectrum. Furthermore, we show that robust training allows one to favor flat areas, as points with large Hessian spectrum show poor robustness to adversarial perturbation. We further study this relationship, and provide empirical and theoretical proof that the inner loop for robust training is a saddle-free optimization problem \textit{almost everywhere}. We present detailed experiments with five different network architectures, including a residual network, tested on MNIST, CIFAR-10, and CIFAR-100 datasets. We have open sourced our method which can be accessed at [1].

cs.CV↗

arXiv2018

On the Computational Inefficiency of Large Batch Sizes for Stochastic Gradient Descent

Increasing the mini-batch size for stochastic gradient descent offers significant opportunities to reduce wall-clock training time, but there are a variety of theoretical and systems challenges that impede the widespread success of this technique. We investigate these issues, with an emphasis on time to convergence and total computational cost, through an extensive empirical analysis of network training across several architectures and problem domains, including image classification, image segmentation, and language modeling. Although it is common practice to increase the batch size in order to fully exploit available computational resources, we find a substantially more nuanced picture. Our main finding is that across a wide range of network architectures and problem domains, increasing the batch size beyond a certain point yields no decrease in wall-clock time to convergence for \emph{either} train or test loss. This batch size is usually substantially below the capacity of current systems. We show that popular training strategies for large batch size optimization begin to fail before we can populate all available compute resources, and we show that the point at which these methods break down depends more on attributes like model architecture and data complexity than it does directly on the size of the dataset.

cs.LG↗

arXiv2018

A Short Introduction to Local Graph Clustering Methods and Software

Graph clustering has many important applications in computing, but due to the increasing sizes of graphs, even traditionally fast clustering methods can be computationally expensive for real-world graphs of interest. Scalability problems led to the development of local graph clustering algorithms that come with a variety of theoretical guarantees. Rather than return a global clustering of the entire graph, local clustering algorithms return a single cluster around a given seed node or set of seed nodes. These algorithms improve scalability because they use time and memory resources that depend only on the size of the cluster returned, instead of the size of the input graph. Indeed, for many of them, their running time grows linearly with the size of the output. In addition to scalability arguments, local graph clustering algorithms have proven to be very useful for identifying and interpreting small-scale and meso-scale structure in large-scale graphs. As opposed to heuristic operational procedures, this class of algorithms comes with strong algorithmic and statistical theory. These include statistical guarantees that prove they have implicit regularization properties. One of the challenges with the existing literature on these approaches is that they are published in a wide variety of areas, including theoretical computer science, statistics, data science, and mathematics. This has made it difficult to relate the various algorithms and ideas together into a cohesive whole. We have recently been working on unifying these diverse perspectives through the lens of optimization as well as providing software to perform these computations in a cohesive fashion. In this note, we provide a brief introduction to local graph clustering, we provide some representative examples of our perspective, and we introduce our software named Local Graph Clustering (LGC).

cs.SI↗