Searcharxiv⌕ Search

arXiv subjects

Junyu Zhang

Publications and source records attributed to Junyu Zhang.

89 records · Page 5Linked to original sources
arXiv2021

SuperNeurons: FFT-based Gradient Sparsification in the Distributed Training of Deep Neural Networks

The performance and efficiency of distributed training of Deep Neural Networks highly depend on the performance of gradient averaging among all participating nodes, which is bounded by the communication between nodes. There are two major strategies to reduce communication overhead: one is to hide communication by overlapping it with computation, and the other is to reduce message sizes. The first solution works well for linear neural architectures, but latest networks such as ResNet and Inception offer limited opportunity for this overlapping. Therefore, researchers have paid more attention to minimizing communication. In this paper, we present a novel gradient compression framework derived from insights of real gradient distributions, and which strikes a balance between compression ratio, accuracy, and computational overhead. Our framework has two major novel components: sparsification of gradients in the frequency domain, and a range-based floating point representation to quantize and further compress gradients frequencies. Both components are dynamic, with tunable parameters that achieve different compression ratio based on the accuracy requirement and systems' platforms, and achieve very high throughput on GPUs. We prove that our techniques guarantee the convergence with a diminishing compression ratio. Our experiments show that the proposed compression framework effectively improves the scalability of most popular neural networks on a 32 GPU cluster to the baseline of no compression, without compromising the accuracy and convergence speed.

cs.DC↗
arXiv2020

Weak Radio Frequency Signal Detection Based on Piezo-Opto-Electro-Mechanical System: Architecture Design and Sensitivity Prediction

We propose a novel radio-frequency (RF) receiving architecture based on micro-electro-mechanical system (MEMS) and optical coherent detection module. The architecture converts the received electrical signal into mechanical vibration through the piezoelectric effect and adopts an optical detection module to detect the mechanical vibration. We analyze the response function of piezoelectric film to an RF signal, the noise limited sensitivity of the optical detection module and the system transfer function in the frequency domain. Finally, we adopt simple on-off keying (OOK) modulation with bandwidth 1 kHz and carrier frequency 1 GHz, to numerically evaluate the detection sensitivity. The result shows that, considering the main noise sources in wireless channel and circuits, the signal detection sensitivity can reach around -160 dBm with a 50 $Ω$ impedance. Such sensitivity significantly outperforms that of the currently deployed Long Term Evolution (LTE) system, when normalizing the transmission bandwidth also to 1 kHz.

eess.SP↗
arXiv2020

Cubic Regularized Newton Method for Saddle Point Models: a Global and Local Convergence Analysis

In this paper, we propose a cubic regularized Newton (CRN) method for solving convex-concave saddle point problems (SPP). At each iteration, a cubic regularized saddle point subproblem is constructed and solved, which provides a search direction for the iterate. With properly chosen stepsizes, the method is shown to converge to the saddle point with global linear and local superlinear convergence rates, if the saddle point function is gradient Lipschitz and strongly-convex-strongly-concave. In the case that the function is merely convex-concave, we propose a homotopy continuation (or path-following) method. Under a Lipschitz-type error bound condition, we present an iteration complexity bound of $\mathcal{O}\left(\ln \left(1/ε\right)\right)$ to reach an $ε$-solution through a homotopy continuation approach, and the iteration complexity bound becomes $\mathcal{O}\left(\left(1/ε\right)^{\frac{1-θ}{θ^2}}\right)$ under a Hölderian-type error bound condition involving a parameter $θ$ ($0<θ<1$).

math.OC↗
arXiv2020

Variational Policy Gradient Method for Reinforcement Learning with General Utilities

In recent years, reinforcement learning (RL) systems with general goals beyond a cumulative sum of rewards have gained traction, such as in constrained problems, exploration, and acting upon prior experiences. In this paper, we consider policy optimization in Markov Decision Problems, where the objective is a general concave utility function of the state-action occupancy measure, which subsumes several of the aforementioned examples as special cases. Such generality invalidates the Bellman equation. As this means that dynamic programming no longer works, we focus on direct policy search. Analogously to the Policy Gradient Theorem \cite{sutton2000policy} available for RL with cumulative rewards, we derive a new Variational Policy Gradient Theorem for RL with general utilities, which establishes that the parametrized policy gradient may be obtained as the solution of a stochastic saddle point problem involving the Fenchel dual of the utility function. We develop a variational Monte Carlo gradient estimation algorithm to compute the policy gradient based on sample paths. We prove that the variational policy gradient scheme converges globally to the optimal policy for the general objective, though the optimization problem is nonconvex. We also establish its rate of convergence of the order $O(1/t)$ by exploiting the hidden convexity of the problem, and proves that it converges exponentially when the problem admits hidden strong convexity. Our analysis applies to the standard RL problem with cumulative rewards as a special case, in which case our result improves the available convergence rate.

cs.LG↗
arXiv2020

Wireless Communication Based on Microwave Photon-Level Detection With Superconducting Devices: Achievable Rate Prediction

Future wireless communication system embraces physical-layer signal detection with high sensitivity, especially in the microwave photon level. Currently, the receiver primarily adopts the signal detection based on semi-conductor devices for signal detection, while this paper introduces high-sensitivity photon-level microwave detection based on superconducting structure. We first overview existing works on the photon-level communication in the optical spectrum as well as the microwave photon-level sensing based on superconducting structure in both theoretical and experimental perspectives, including microwave detection circuit model based on Josephson junction, microwave photon counter based on Josephson junction, and two reconstruction approaches under background noise. In addition, we characterize channel modeling based on two different microwave photon detection approaches, including the absorption barrier and the dual-path Handury Brown-Twiss (HBT) experiments, and predict the corresponding achievable rates. According to the performance prediction, it is seen that the microwave photon-level signal detection can increase the receiver sensitivity compared with the state-of-the-art standardized communication system with waveform signal reception, with gain over $10$dB.

eess.SP↗
arXiv2020

On the Divergence of Decentralized Non-Convex Optimization

We study a generic class of decentralized algorithms in which $N$ agents jointly optimize the non-convex objective $f(u):=1/N\sum_{i=1}^{N}f_i(u)$, while only communicating with their neighbors. This class of problems has become popular in modeling many signal processing and machine learning applications, and many efficient algorithms have been proposed. However, by constructing some counter-examples, we show that when certain local Lipschitz conditions (LLC) on the local function gradient $\nabla f_i$'s are not satisfied, most of the existing decentralized algorithms diverge, even if the global Lipschitz condition (GLC) is satisfied, where the sum function $f$ has Lipschitz gradient. This observation raises an important open question: How to design decentralized algorithms when the LLC, or even the GLC, is not satisfied? To address the above question, we design a first-order algorithm called Multi-stage gradient tracking algorithm (MAGENTA), which is capable of computing stationary solutions with neither the LLC nor the GLC. In particular, we show that the proposed algorithm converges sublinearly to certain $ε$-stationary solution, where the precise rate depends on various algorithmic and problem parameters. In particular, if the local function $f_i$'s are $Q$th order polynomials, then the rate becomes $\mathcal{O}(1/ε^{Q-1})$. Such a rate is tight for the special case of $Q=2$ where each $f_i$ satisfies LLC. To our knowledge, this is the first attempt that studies decentralized non-convex optimization problems with neither the LLC nor the GLC.

math.OC↗
arXiv2020

Generalization Bounds for Stochastic Saddle Point Problems

This paper studies the generalization bounds for the empirical saddle point (ESP) solution to stochastic saddle point (SSP) problems. For SSP with Lipschitz continuous and strongly convex-strongly concave objective functions, we establish an $\mathcal{O}(1/n)$ generalization bound by using a uniform stability argument. We also provide generalization bounds under a variety of assumptions, including the cases without strong convexity and without bounded domains. We illustrate our results in two examples: batch policy learning in Markov decision process, and mixed strategy Nash equilibrium estimation for stochastic games. In each of these examples, we show that a regularized ESP solution enjoys a near-optimal sample complexity. To the best of our knowledge, this is the first set of results on the generalization theory of ESP.

math.OC↗
arXiv2020

Finite-size analysis of continuous variable source-independent quantum random number generation

We study the impact of finite-size effect on continuous variable source-independent quantum random number generation. The central-limit theorem and maximum likelihood estimation theorem are used to derive the formula which could output the statistical fluctuations and determine upper bound of parameters of practical quantum random number generation. With these results, we can see the check data length and confidence probability has intense relevance to the final randomness, which can be adjusted according to the demand in implementation. Besides, other key parameters, such as sampling range size and sampling resolution, have also been considered in detail. It is found that the distribution of quantified output related with sampling range size has significant effects on the loss of final randomness due to finite-size effect. The overall results indicate that the finite-size effect should be taken into consideration for implementing the continuous variable source-independent quantum random number generation in practical.

quant-ph↗
arXiv2020

Cautious Reinforcement Learning via Distributional Risk in the Dual Domain

We study the estimation of risk-sensitive policies in reinforcement learning problems defined by a Markov Decision Process (MDPs) whose state and action spaces are countably finite. Prior efforts are predominately afflicted by computational challenges associated with the fact that risk-sensitive MDPs are time-inconsistent. To ameliorate this issue, we propose a new definition of risk, which we call caution, as a penalty function added to the dual objective of the linear programming (LP) formulation of reinforcement learning. The caution measures the distributional risk of a policy, which is a function of the policy's long-term state occupancy distribution. To solve this problem in an online model-free manner, we propose a stochastic variant of primal-dual method that uses Kullback-Lieber (KL) divergence as its proximal term. We establish that the number of iterations/samples required to attain approximately optimal solutions of this scheme matches tight dependencies on the cardinality of the state and action spaces, but differs in its dependence on the infinity norm of the gradient of the risk measure. Experiments demonstrate the merits of this approach for improving the reliability of reward accumulation without additional computational burdens.

stat.ML↗
arXiv2019

From low probability to high confidence in stochastic convex optimization

Standard results in stochastic convex optimization bound the number of samples that an algorithm needs to generate a point with small function value in expectation. More nuanced high probability guarantees are rare, and typically either rely on "light-tail" noise assumptions or exhibit worse sample complexity. In this work, we show that a wide class of stochastic optimization algorithms for strongly convex problems can be augmented with high confidence bounds at an overhead cost that is only logarithmic in the confidence level and polylogarithmic in the condition number. The procedure we propose, called proxBoost, is elementary and builds on two well-known ingredients: robust distance estimation and the proximal point method. We discuss consequences for both streaming (online) algorithms and offline algorithms based on empirical risk minimization.

math.OC↗
arXiv2019

A Stochastic Composite Gradient Method with Incremental Variance Reduction

We consider the problem of minimizing the composition of a smooth (nonconvex) function and a smooth vector mapping, where the inner mapping is in the form of an expectation over some random variable or a finite sum. We propose a stochastic composite gradient method that employs an incremental variance-reduced estimator for both the inner vector mapping and its Jacobian. We show that this method achieves the same orders of complexity as the best known first-order methods for minimizing expected-value and finite-sum nonconvex functions, despite the additional outer composition which renders the composite gradient estimator biased. This finding enables a much broader range of applications in machine learning to benefit from the low complexity of incremental variance-reduction methods.

math.OC↗
arXiv2018

Adaptive Stochastic Variance Reduction for Subsampled Newton Method with Cubic Regularization

The cubic regularized Newton method of Nesterov and Polyak has become increasingly popular for non-convex optimization because of its capability of finding an approximate local solution with second-order guarantee. Several recent works extended this method to the setting of minimizing the average of N smooth functions by replacing the exact gradients and Hessians with subsampled approximations. It has been shown that the total Hessian sample complexity can be reduced to be sublinear in N per iteration by leveraging stochastic variance reduction techniques. We present an adaptive variance reduction scheme for subsampled Newton method with cubic regularization, and show that the expected Hessian sample complexity is O(N + N^{2/3}ε^{-3/2}) for finding an (ε,ε^{1/2})-approximate local solution (in terms of first and second-order guarantees respectively). Moreover, we show that the same Hessian sample complexity retains with fixed sample sizes if exact gradients are used. The techniques of our analysis are different from previous works in that we do not rely on high probability bounds based on matrix concentration inequalities. Instead, we derive and utilize bounds on the 3rd and 4th order moments of the average of random matrices, which are of independent interest on their own.

math.OC↗
arXiv2018

A Cubic Regularized Newton's Method over Riemannian Manifolds

In this paper we present a cubic regularized Newton's method to minimize a smooth function over a Riemannian manifold. The proposed algorithm is shown to reach a second-order $ε$-stationary point within $\mathcal{O}(1/ε^{\frac{3}{2}})$ iterations, under the condition that the pullbacks are locally Lipschitz continuous, a condition that is shown to be satisfied if the manifold is compact. Furthermore, we present a local superlinear convergence result under some additional conditions.

math.OC↗
arXiv2018

Highly accurate model for prediction of lung nodule malignancy with CT scans

Computed tomography (CT) examinations are commonly used to predict lung nodule malignancy in patients, which are shown to improve noninvasive early diagnosis of lung cancer. It remains challenging for computational approaches to achieve performance comparable to experienced radiologists. Here we present NoduleX, a systematic approach to predict lung nodule malignancy from CT data, based on deep learning convolutional neural networks (CNN). For training and validation, we analyze >1000 lung nodules in images from the LIDC/IDRI cohort. All nodules were identified and classified by four experienced thoracic radiologists who participated in the LIDC project. NoduleX achieves high accuracy for nodule malignancy classification, with an AUC of ~0.99. This is commensurate with the analysis of the dataset by experienced radiologists. Our approach, NoduleX, provides an effective framework for highly accurate nodule malignancy prediction with the model trained on a large patient population. Our results are replicable with software available at http://bioinformatics.astate.edu/NoduleX.

cs.CV↗
arXiv2017

Primal-Dual Optimization Algorithms over Riemannian Manifolds: an Iteration Complexity Analysis

In this paper we study nonconvex and nonsmooth multi-block optimization over Riemannian manifolds with coupled linear constraints. Such optimization problems naturally arise from machine learning, statistical learning, compressive sensing, image processing, and tensor PCA, among others. We develop an ADMM-like primal-dual approach based on decoupled solvable subroutines such as linearized proximal mappings. First, we introduce the optimality conditions for the afore-mentioned optimization models. Then, the notion of $ε$-stationary solutions is introduced as a result. The main part of the paper is to show that the proposed algorithms enjoy an iteration complexity of $O(1/ε^2)$ to reach an $ε$-stationary solution. For prohibitively large-size tensor or machine learning models, we present a sampling-based stochastic algorithm with the same iteration complexity bound in expectation. In case the subproblems are not analytically solvable, a feasible curvilinear line-search variant of the algorithm based on retraction operators is proposed. Finally, we show specifically how the algorithms can be implemented to solve a variety of practical problems such as the NP-hard maximum bisection problem, the $\ell_q$ regularized sparse tensor principal component analysis and the community detection problem. Our preliminary numerical results show great potentials of the proposed methods.

math.OC↗
arXiv2017

A Sparse Completely Positive Relaxation of the Modularity Maximization for Community Detection

In this paper, we consider the community detection problem under either the stochastic block model (SBM) assumption or the degree-correlated stochastic block model (DCSBM) assumption. The modularity maximization formulation for the community detection problem is NP-hard in general. In this paper, we propose a sparse and low-rank completely positive relaxation for the modularity maximization problem, we then develop an efficient row-by-row (RBR) type block coordinate descent (BCD) algorithm to solve the relaxation and prove an $\mathcal{O}(1/\sqrt{N})$ convergence rate to a stationary point where $N$ is the number of iterations. A fast rounding scheme is constructed to retrieve the community structure from the solution. Non-asymptotic high probability bounds on the misclassification rate are established to justify our approach. We further develop an asynchronous parallel RBR algorithm to speed up the convergence. Extensive numerical experiments on both synthetic and real world networks show that the proposed approach enjoys advantages in both clustering accuracy and numerical efficiency. Our numerical results indicate that the newly proposed method is a quite competitive alternative for community detection on sparse networks with over 50 million nodes.

math.OC↗