arXiv subjects

Zhihua Zhang

Publications and source records attributed to Zhihua Zhang.

At least 91 records · Page 5Linked to original sources

arXiv2021

Communication-Efficient Distributed SVD via Local Power Iterations

We study distributed computing of the truncated singular value decomposition problem. We develop an algorithm that we call \texttt{LocalPower} for improving communication efficiency. Specifically, we uniformly partition the dataset among $m$ nodes and alternate between multiple (precisely $p$) local power iterations and one global aggregation. In the aggregation, we propose to weight each local eigenvector matrix with orthogonal Procrustes transformation (OPT). As a practical surrogate of OPT, sign-fixing, which uses a diagonal matrix with $\pm 1$ entries as weights, has better computation complexity and stability in experiments. We theoretically show that under certain assumptions \texttt{LocalPower} lowers the required number of communications by a factor of $p$ to reach a constant accuracy. We also show that the strategy of periodically decaying $p$ helps obtain high-precision solutions. We conduct experiments to demonstrate the effectiveness of \texttt{LocalPower}.

stat.ML↗

arXiv2021

Explicit Superlinear Convergence Rates of The SR1 Algorithm

We study the convergence rate of the famous Symmetric Rank-1 (SR1) algorithm which has wide applications in different scenarios. Although it has been extensively investigated, SR1 still lacks a non-asymptotic superlinear rate compared with other quasi-Newton methods such as DFP and BFGS. In this paper we address this problem. Inspired by the recent work on explicit convergence analysis of quasi-Newton methods, we obtain the first explicit non-asymptotic rates of superlinear convergence for the vanilla SR1 methods with correction strategy to achieve the numerical stability. Specifically, the vanilla SR1 with the correction strategy achieves the rates of the form $\left(\frac{4n\ln(eκ) }{k}\right)^{k/2}$ for general smooth strongly-convex functions where $k$ is the iteration counter, $κ$ is the condition number of the objective function and $n$ is the dimension of the problem. For the quadratic function, the vanilla SR1 algorithm can find the optima of the objective function at most $n$ steps.

math.OC↗

arXiv2021

Memory-Efficient Differentiable Transformer Architecture Search

Differentiable architecture search (DARTS) is successfully applied in many vision tasks. However, directly using DARTS for Transformers is memory-intensive, which renders the search process infeasible. To this end, we propose a multi-split reversible network and combine it with DARTS. Specifically, we devise a backpropagation-with-reconstruction algorithm so that we only need to store the last layer's outputs. By relieving the memory burden for DARTS, it allows us to search with larger hidden size and more candidate operations. We evaluate the searched architecture on three sequence-to-sequence datasets, i.e., WMT'14 English-German, WMT'14 English-French, and WMT'14 English-Czech. Experimental results show that our network consistently outperforms standard Transformers across the tasks. Moreover, our method compares favorably with big-size Evolved Transformers, reducing search computation by an order of magnitude.

cs.LG↗

arXiv2021

Directional Convergence Analysis under Spherically Symmetric Distribution

We consider the fundamental problem of learning linear predictors (i.e., separable datasets with zero margin) using neural networks with gradient flow or gradient descent. Under the assumption of spherically symmetric data distribution, we show directional convergence guarantees with exact convergence rate for two-layer non-linear networks with only two hidden nodes, and (deep) linear networks. Moreover, our discovery is built on dynamic from the initialization without both initial loss and perfect classification constraint in contrast to previous works. We also point out and study the challenges in further strengthening and generalizing our results.

cs.LG↗

arXiv2021

Meta-Regularization: An Approach to Adaptive Choice of the Learning Rate in Gradient Descent

We propose \textit{Meta-Regularization}, a novel approach for the adaptive choice of the learning rate in first-order gradient descent methods. Our approach modifies the objective function by adding a regularization term on the learning rate, and casts the joint updating process of parameters and learning rates into a maxmin problem. Given any regularization term, our approach facilitates the generation of practical algorithms. When \textit{Meta-Regularization} takes the $φ$-divergence as a regularizer, the resulting algorithms exhibit comparable theoretical convergence performance with other first-order gradient-based algorithms. Furthermore, we theoretically prove that some well-designed regularizers can improve the convergence performance under the strong-convexity condition of the objective function. Numerical experiments on benchmark problems demonstrate the effectiveness of algorithms derived from some common $φ$-divergence in full batch as well as online learning settings.

cs.LG↗

arXiv2021

Communication-Efficient Local Decentralized SGD Methods

Recently, the technique of local updates is a powerful tool in centralized settings to improve communication efficiency via periodical communication. For decentralized settings, it is still unclear how to efficiently combine local updates and decentralized communication. In this work, we propose an algorithm named as LD-SGD, which incorporates arbitrary update schemes that alternate between multiple Local updates and multiple Decentralized SGDs, and provide an analytical framework for LD-SGD. Under the framework, we present a sufficient condition to guarantee the convergence. We show that LD-SGD converges to a critical point for a wide range of update schemes when the objective is non-convex and the training data are non-identically independent distributed. Moreover, our framework brings many insights into the design of update schemes for decentralized optimization. As examples, we specify two update schemes and show how they help improve communication efficiency. Specifically, the first scheme alternates the number of local and global update steps. From our analysis, the ratio of the number of local updates to that of decentralized SGD trades off communication and computation. The second scheme is to periodically shrink the length of local updates. We show that the decaying strategy helps improve communication efficiency both theoretically and empirically.

stat.ML↗

arXiv2021

DIPPA: An improved Method for Bilinear Saddle Point Problems

This paper studies bilinear saddle point problems $\min_{\bf{x}} \max_{\bf{y}} g(\bf{x}) + \bf{x}^{\top} \bf{A} \bf{y} - h(\bf{y})$, where the functions $g, h$ are smooth and strongly-convex. When the gradient and proximal oracle related to $g$ and $h$ are accessible, optimal algorithms have already been developed in the literature \cite{chambolle2011first, palaniappan2016stochastic}. However, the proximal operator is not always easy to compute, especially in constraint zero-sum matrix games \cite{zhang2020sparsified}. This work proposes a new algorithm which only requires the access to the gradients of $g, h$. Our algorithm achieves a complexity upper bound $\tilde{\mathcal{O}}\left( \frac{\|\bf{A}\|_2}{\sqrt{μ_x μ_y}} + \sqrt[4]{κ_x κ_y (κ_x + κ_y)} \right)$ which has optimal dependency on the coupling condition number $\frac{\|\bf{A}\|_2}{\sqrt{μ_x μ_y}}$ up to logarithmic factors.

cs.LG↗

arXiv2021

Delayed Projection Techniques for Linearly Constrained Problems: Convergence Rates, Acceleration, and Applications

In this work, we study a novel class of projection-based algorithms for linearly constrained problems (LCPs) which have a lot of applications in statistics, optimization, and machine learning. Conventional primal gradient-based methods for LCPs call a projection after each (stochastic) gradient descent, resulting in that the required number of projections equals that of gradient descents (or total iterations). Motivated by the recent progress in distributed optimization, we propose the delayed projection technique that calls a projection once for a while, lowering the projection frequency and improving the projection efficiency. Accordingly, we devise a series of stochastic methods for LCPs using the technique, including a variance reduced method and an accelerated one. We theoretically show that it is feasible to improve projection efficiency in both strongly convex and generally convex cases. Our analysis is simple and unified and can be easily extended to other methods using delayed projections. When applying our new algorithms to federated optimization, a newfangled and privacy-preserving subfield in distributed optimization, we obtain not only a variance reduced federated algorithm with convergence rates better than previous works, but also the first accelerated method able to handle data heterogeneity inherent in federated optimization.

math.OC↗

arXiv2020

An Asymptotically Optimal Multi-Armed Bandit Algorithm and Hyperparameter Optimization

The evaluation of hyperparameters, neural architectures, or data augmentation policies becomes a critical model selection problem in advanced deep learning with a large hyperparameter search space. In this paper, we propose an efficient and robust bandit-based algorithm called Sub-Sampling (SS) in the scenario of hyperparameter search evaluation. It evaluates the potential of hyperparameters by the sub-samples of observations and is theoretically proved to be optimal under the criterion of cumulative regret. We further combine SS with Bayesian Optimization and develop a novel hyperparameter optimization algorithm called BOSS. Empirical studies validate our theoretical arguments of SS and demonstrate the superior performance of BOSS on a number of applications, including Neural Architecture Search (NAS), Data Augmentation (DA), Object Detection (OD), and Reinforcement Learning (RL).

stat.ML↗

arXiv2020

Improved simulation of El Niño and its influence on the climate anomalies of the East Asia-western North Pacific in the ICM Version 2

This study introduces the second version of the Integrated Climate Model (ICM). ICM is developed by the Center for Monsoon System Research, Institute of Atmospheric Physics to improve the short-term climate prediction of the East Asia-western North Pacific (EA-WNP). The main update of the second version of ICM (ICM.V2) relative to the first version (ICM.V1) is the improvement of the horizontal resolution of the atmospheric model from T31 spectral resolution (3.75°*3.75°) to T63 (1.875°*1.875°). As a result, some important factors for the short-term climate prediction of the EA-WNP is apparently improved from ICM.V1 to ICM.V2, including the climatological SST, the rainfall and circulation of the East Asian summer monsoon, and the variability and spatial pattern of ENSO. The impact of El Niño on the EA-WNP climate simulated in ICM.V2 is also improved with more realistic anticyclonic anomalies and precipitation pattern over the EA-WNP. The tropical Indian ocean capacitor effect and the WNP local air-sea interaction feedback, two popular mechanisms to explain the impact of El Niño on the EA-WNP climate is also realistically reproduced in ICM.V2, much improved relative to that in ICM.V1.

physics.ao-ph↗

arXiv2020

Finding the Near Optimal Policy via Adaptive Reduced Regularization in MDPs

Regularized MDPs serve as a smooth version of original MDPs. However, biased optimal policy always exists for regularized MDPs. Instead of making the coefficientλof regularized term sufficiently small, we propose an adaptive reduction scheme for λ to approximate optimal policy of the original MDP. It is shown that the iteration complexity for obtaining anε-optimal policy could be reduced in comparison with setting sufficiently smallλ. In addition, there exists strong duality connection between the reduction method and solving the original MDP directly, from which we can derive more adaptive reduction method for certain algorithms.

cs.LG↗

arXiv2020

Optimal Quantization for Batch Normalization in Neural Network Deployments and Beyond

Quantized Neural Networks (QNNs) use low bit-width fixed-point numbers for representing weight parameters and activations, and are often used in real-world applications due to their saving of computation resources and reproducibility of results. Batch Normalization (BN) poses a challenge for QNNs for requiring floating points in reciprocal operations, and previous QNNs either require computing BN at high precision or revise BN to some variants in heuristic ways. In this work, we propose a novel method to quantize BN by converting an affine transformation of two floating points to a fixed-point operation with shared quantized scale, which is friendly for hardware acceleration and model deployment. We confirm that our method maintains same outputs through rigorous theoretical analysis and numerical analysis. Accuracy and efficiency of our quantization method are verified by experiments at layer level on CIFAR and ImageNet datasets. We also believe that our method is potentially useful in other problems involving quantization.

cs.LG↗

arXiv2020

Intervention Generative Adversarial Networks

In this paper we propose a novel approach for stabilizing the training process of Generative Adversarial Networks as well as alleviating the mode collapse problem. The main idea is to introduce a regularization term that we call intervention loss into the objective. We refer to the resulting generative model as Intervention Generative Adversarial Networks (IVGAN). By perturbing the latent representations of real images obtained from an auxiliary encoder network with Gaussian invariant interventions and penalizing the dissimilarity of the distributions of the resulting generated images, the intervention loss provides more informative gradient for the generator, significantly improving GAN's training stability. We demonstrate the effectiveness and efficiency of our methods via solid theoretical analysis and thorough evaluation on standard real-world datasets as well as the stacked MNIST dataset.

stat.ML↗

arXiv2020

On the Convergence of FedAvg on Non-IID Data

Federated learning enables a large amount of edge computing devices to jointly learn a model without data sharing. As a leading algorithm in this setting, Federated Averaging (\texttt{FedAvg}) runs Stochastic Gradient Descent (SGD) in parallel on a small subset of the total devices and averages the sequences only once in a while. Despite its simplicity, it lacks theoretical guarantees under realistic settings. In this paper, we analyze the convergence of \texttt{FedAvg} on non-iid data and establish a convergence rate of $\mathcal{O}(\frac{1}{T})$ for strongly convex and smooth problems, where $T$ is the number of SGDs. Importantly, our bound demonstrates a trade-off between communication-efficiency and convergence rate. As user devices may be disconnected from the server, we relax the assumption of full device participation to partial device participation and study different averaging schemes; low device participation rate can be achieved without severely slowing down the learning. Our results indicate that heterogeneity of data slows down the convergence, which matches empirical observations. Furthermore, we provide a necessary condition for \texttt{FedAvg} on non-iid data: the learning rate $η$ must decay, even if full-gradient is used; otherwise, the solution will be $Ω(η)$ away from the optimal.

stat.ML↗

arXiv2020

Approximate Newton Methods

Many machine learning models involve solving optimization problems. Thus, it is important to deal with a large-scale optimization problem in big data applications. Recently, subsampled Newton methods have emerged to attract much attention due to their efficiency at each iteration, rectified a weakness in the ordinary Newton method of suffering a high cost in each iteration while commanding a high convergence rate. Other efficient stochastic second order methods are also proposed. However, the convergence properties of these methods are still not well understood. There are also several important gaps between the current convergence theory and the performance in real applications. In this paper, we aim to fill these gaps. We propose a unifying framework to analyze both local and global convergence properties of second order methods. Based on this framework, we present our theoretical results which match the performance in real applications well.

math.NA↗

arXiv2019

Fast Generalized Matrix Regression with Applications in Machine Learning

Fast matrix algorithms have become the fundamental tools of machine learning in big data era. The generalized matrix regression problem is widely used in the matrix approximation such as CUR decomposition, kernel matrix approximation, and stream singular value decomposition (SVD), etc. In this paper, we propose a fast generalized matrix regression algorithm (Fast GMR) which utilizes sketching technique to solve the GMR problem efficiently. Given error parameter $0<ε<1$, the Fast GMR algorithm can achieve a $(1+ε)$ relative error with the sketching sizes being of order $\cO(ε^{-1/2})$ for a large group of GMR problems. We apply the Fast GMR algorithm to the symmetric positive definite matrix approximation and single pass singular value decomposition and they achieve a better performance than conventional algorithms. Our empirical study also validates the effectiveness and efficiency of our proposed algorithms.

cs.LG↗

arXiv2019

A Regularized Approach to Sparse Optimal Policy in Reinforcement Learning

We propose and study a general framework for regularized Markov decision processes (MDPs) where the goal is to find an optimal policy that maximizes the expected discounted total reward plus a policy regularization term. The extant entropy-regularized MDPs can be cast into our framework. Moreover, under our framework, many regularization terms can bring multi-modality and sparsity, which are potentially useful in reinforcement learning. In particular, we present sufficient and necessary conditions that induce a sparse optimal policy. We also conduct a full mathematical analysis of the proposed regularized MDPs, including the optimality condition, performance error, and sparseness control. We provide a generic method to devise regularization forms and propose off-policy actor critic algorithms in complex environment settings. We empirically analyze the numerical properties of optimal policies and compare the performance of different sparse regularization forms in discrete and continuous environments.

stat.ML↗

arXiv2019

Distillation $\approx$ Early Stopping? Harvesting Dark Knowledge Utilizing Anisotropic Information Retrieval For Overparameterized Neural Network

Distillation is a method to transfer knowledge from one model to another and often achieves higher accuracy with the same capacity. In this paper, we aim to provide a theoretical understanding on what mainly helps with the distillation. Our answer is "early stopping". Assuming that the teacher network is overparameterized, we argue that the teacher network is essentially harvesting dark knowledge from the data via early stopping. This can be justified by a new concept, {Anisotropic Information Retrieval (AIR)}, which means that the neural network tends to fit the informative information first and the non-informative information (including noise) later. Motivated by the recent development on theoretically analyzing overparameterized neural networks, we can characterize AIR by the eigenspace of the Neural Tangent Kernel(NTK). AIR facilities a new understanding of distillation. With that, we further utilize distillation to refine noisy labels. We propose a self-distillation algorithm to sequentially distill knowledge from the network in the previous training epoch to avoid memorizing the wrong labels. We also demonstrate, both theoretically and empirically, that self-distillation can benefit from more than just early stopping. Theoretically, we prove convergence of the proposed algorithm to the ground truth labels for randomly initialized overparameterized neural networks in terms of $\ell_2$ distance, while the previous result was on convergence in $0$-$1$ loss. The theoretical result ensures the learned neural network enjoy a margin on the training data which leads to better generalization. Empirically, we achieve better testing accuracy and entirely avoid early stopping which makes the algorithm more user-friendly.

stat.ML↗