Searcharxiv⌕ Search

arXiv subjects

Wei Peng

Publications and source records attributed to Wei Peng.

198 records · Page 11Linked to original sources
arXiv2019

Training GANs with Centripetal Acceleration

Training generative adversarial networks (GANs) often suffers from cyclic behaviors of iterates. Based on a simple intuition that the direction of centripetal acceleration of an object moving in uniform circular motion is toward the center of the circle, we present the Simultaneous Centripetal Acceleration (SCA) method and the Alternating Centripetal Acceleration (ACA) method to alleviate the cyclic behaviors. Under suitable conditions, gradient descent methods with either SCA or ACA are shown to be linearly convergent for bilinear games. Numerical experiments are conducted by applying ACA to existing gradient-based algorithms in a GAN setup scenario, which demonstrate the superiority of ACA.

cs.LG↗
arXiv2019

A Boost in Revealing Subtle Facial Expressions: A Consolidated Eulerian Framework

Facial Micro-expression Recognition (MER) distinguishes the underlying emotional states of spontaneous subtle facialexpressions. Automatic MER is challenging because that 1) the intensity of subtle facial muscle movement is extremely lowand 2) the duration of ME is transient.Recent works adopt motion magnification or time interpolation to resolve these issues. Nevertheless, existing works dividethem into two separate modules due to their non-linearity. Though such operation eases the difficulty in implementation, itignores their underlying connections and thus results in inevitable losses in both accuracy and speed. Instead, in this paper, weexplore their underlying joint formulations and propose a consolidated Eulerian framework to reveal the subtle facial movements.It expands the temporal duration and amplifies the muscle movements in micro-expressions simultaneously. Compared toexisting approaches, the proposed method can not only process ME clips more efficiently but also make subtle ME movementsmore distinguishable. Experiments on two public MER databases indicate that our model outperforms the state-of-the-art inboth speed and accuracy.

cs.CV↗
arXiv2018

Global Complexity Analysis of Inexact Successive Quadratic Approximation methods for Regularized Optimization under Mild Assumptions

Successive quadratic approximations (SQA) are numerically efficient for minimizing the sum of a smooth function and a convex function. The iteration complexity of inexact SQA methods has been analyzed recently. In this paper, we present an algorithmic framework of inexact SQA methods with four types of line searches, and analyze its global complexity under milder assumptions. First, we show its well-definedness and some decreasing properties. Second, under the quadratic growth condition and a uniform positive lower bound condition on stepsizes, we show that the function value sequence and the iterate sequence are linearly convergent. Moreover, we obtain a o(1/k) complexity without the quadratic growth condition, improving existing O(1/k) complexity results. At last, we show that a local gradient-Lipschitz-continuity condition could guarantee a uniform positive lower bound for the stepsizes.

math.OC↗
arXiv2018

A linear time algorithm for multiscale quantile simulation

Change-point problems have appeared in a great many applications for example cancer genetics, econometrics and climate change. Modern multiscale type segmentation methods are considered to be a statistically efficient approach for multiple change-point detection, which minimize the number of change-points under a multiscale side-constraint. The constraint threshold plays a critical role in balancing the data-fit and model complexity. However, the computation time of such a threshold is quadratic in terms of sample size $n$, making it impractical for large scale problems. In this paper we proposed an $\mathcal{O}(n)$ algorithm by utilizing the hidden quasiconvexity structure of the problem. It applies to all regression models in exponential family with arbitrary convex scale penalties. Simulations verify its computational efficiency and accuracy. An implementation is provided in R-package "linearQ" on CRAN.

stat.CO↗
arXiv2018

Nonconvex Proximal Incremental Aggregated Gradient Method with Linear Convergence

In this paper, we study the proximal incremental aggregated gradient(PIAG) algorithm for minimizing the sum of L-smooth nonconvex component functions and a proper closed convex function. By exploiting the L-smooth property and with the help of an error bound condition, we can show that the PIAG method still enjoys some nice linear convergence properties even for nonconvex minimization. To illustrate this, we first demonstrate that the generated sequence globally converges to the stationary point set. Then, there exists a threshold such that the objective function value sequence and the iterate point sequence are R-linearly convergent when the stepsize is chosen below this threshold.

math.OC↗
arXiv2017

Inertial Proximal Incremental Aggregated Gradient Method

In this paper, we introduce an inertial version of the Proximal Incremental Aggregated Gradient method (PIAG) for minimizing the sum of smooth convex component functions and a possibly nonsmooth convex regularization function. Theoretically, we show that the inertial Proximal Incremental Aggregated Gradiend (iPIAG) method enjoys a global linear convergence under a quadratic growth condition, which is strictly weaker than strong convexity, provided that the stepsize is not larger than a constant. Moreover, we present two numerical expreiments which demonstrate that iPIAG outperforms the original PIAG.

math.OC↗
arXiv2017

Ultra-Dense HetNets Meet Big Data: Green Frameworks, Techniques, and Approaches

Ultra-dense heterogeneous networks (Ud-HetNets) have been put forward to improve the network capacity for next-generation wireless networks. However, counter to the 5G vision, ultra-dense deployment of networks would significantly increase energy consumption and thus decrease network energy efficiency suffering from the conventional worst-case network design philosophy. This problem becomes particularly severe when Ud-HetNets meet big data because of the traditional reactive request-transmit service mode. In view of these, this article first develops a big-data-aware artificial intelligent based framework for energy-efficient operations of Ud-HetNets. Based on the framework, we then identify four promising techniques, namely big data analysis, adaptive base station operation, proactive caching, and interference-aware resource allocation, to reduce energy cost on both large and small scales. We further develop a load-aware stochastic optimization approach to show the potential of our proposed framework and techniques in energy conservation. In a nutshell, we devote to constructing green Ud-HetNets of big data with the abilities of learning and inferring by improving the flexibility of control from worst-case to adaptive design and shifting the manner of services from reactive to proactive modes.

cs.IT↗
arXiv2014

Nonideal optical cavity structure of superconducting nanowire single photon detector

Optical cavity structure has been proven to be a crucial factor for obtaining high detection efficiency in superconducting nanowire single photon detector (SNSPD). Practically, complicated fabrication processes may result in a non-ideal optical cavity structure. The cross-sectional transmission electron microscope (TEM) image of SNSPD fabricated in this study shows unexpected arc-shaped optical cavities which could have originated due to the over-etching of SiO2 layer while defining NbN nanowire. The effects of the arc-shaped optical cavity structure, such as the wavelength dependence of the optical absorption efficiency for different polarization, were analyzed by performing optical simulations using finite-difference time-domain method. The central wavelength of the device is found to exhibit a blue shift owing to the arced cavity structure. This effect is equivalent to the flat cavity with a reduced height. The results may give interesting reference for SNSPD design and fabrication.

cond-mat.supr-con↗
arXiv2014

Superconducting nanowire single photon detector with on-chip bandpass filter

Dark count rate is one of the key parameters limiting the performance of the superconducting nanowire single photon detector (SNSPD). We have designed a multi-layer film bandpass filter that can be integrated onto the SNSPD to suppress the dark counts contributed by the stray light and blackbody radiation of the fiber. The bandpass filter is composed of 16 SiO2/Si bilayers deposited onto the backside of a thermally oxidized Si substrate. The substrate shows an excellent bandpass filter effect and provides a high transmittance of 88% at the central wavelength of the pass band, which is the same as that of the bare substrate. The SNSPDs fabricated on the substrate integrated with the bandpass filter show conspicuous wavelength-sensitive detection efficiency. The background dark count rate is reduced by two orders of magnitude to sub-Hz compared with the conventional SNSPD (a few tens of Hz). The detector exhibits a system detection efficiency of 56% at DCR of 1 Hz, with the measured minimal noise equivalent power reaching 2e-19 w/Hz1/2.

cond-mat.supr-con↗
arXiv2013

Sparse Channel Estimation for Amplify-and-Forward Two-way Relay Network with Compressed Sensing

Amplify-and-forward two-way relay network (AFTWRN) was introduced to realize high-data rate transmission over the wireless frequency-selective channel. However, AFTWRC requires the knowledge of channel state information (CSI) not only for coherent data detection but also for the selfdata removal. This is partial accomplished by training sequence-based linear channel estimation. However, conventional linear estimation techniques neglect anticipated sparsity of multipath channel and thus lead to low spectral efficiency which is scarce in the field of wireless communication. Unlike the previous methods, we propose a sparse channel estimation method which can exploit the sparse structure and hence provide significant improvements in MSE performance when compared with traditional LS-based linear channel probing strategies in AF-TWRN. Simulation results confirm the proposed methods.

cs.IT↗
arXiv2013

Adaptive Sparse Channel Estimation for Time-Variant MIMO-OFDM Systems

Accurate channel state information (CSI) is required for coherent detection in time-variant multiple-input multipleoutput (MIMO) communication systems using orthogonal frequency division multiplexing (OFDM) modulation. One of low-complexity and stable adaptive channel estimation (ACE) approaches is the normalized least mean square (NLMS)-based ACE. However, it cannot exploit the inherent sparsity of MIMO channel which is characterized by a few dominant channel taps. In this paper, we propose two adaptive sparse channel estimation (ASCE) methods to take advantage of such sparse structure information for time-variant MIMO-OFDM systems. Unlike traditional NLMS-based method, two proposed methods are implemented by introducing sparse penalties to the cost function of NLMS algorithm. Computer simulations confirm obvious performance advantages of the proposed ASCEs over the traditional ACE.

cs.IT↗
arXiv2013

Adaptive Sparse Channel Estimation for Time-Variant MISO Communication Systems

Channel estimation problem is one of the key technical issues in time-variant multiple-input single-output (MSIO) communication systems. To estimate the MISO channel, least mean square (LMS) algorithm is applied to adaptive channel estimation (ACE). Since the MISO channel is often described by sparse channel model, such sparsity can be exploited and then estimation performance can be improved by adaptive sparse channel estimation (ASCE) methods using sparse LMS algorithms. However, conventional ASCE methods have two main drawbacks: 1) sensitive to random scale of training signal and 2) unstable in low signal-to-noise ratio (SNR) regime. To overcome these two harmful factors, in this paper, we propose a novel ASCE method using normalized LMS (NLMS) algorithm (ASCE-NLMS). In addition, we also proposed an improved ASCE method using normalized least mean fourth (NLMF) algorithm (ASCE-NLMF). Two proposed methods can exploit the channel sparsity effectively. Also, stability of the proposed methods is confirmed by mathematical derivation. Computer simulation results show that the proposed sparse channel estimation methods can achieve better estimation performance than conventional methods.

cs.IT↗
arXiv2013

Sparse Channel Estimation for MIMO-OFDM Amplify-and-Forward Two-Way Relay Networks

Accurate channel impulse response (CIR) is required for coherent detection and it can also help improve communication quality of service in next-generation wireless communication systems. One of the advanced systems is multi-input multi-output orthogonal frequency-division multiplexing (MIMO-OFDM) amplify and forward two-way relay networks (AF-TWRN). Linear channel estimation methods, e.g., least square (LS), have been proposed to estimate the CIR. However, these methods never take advantage of channel sparsity and then cause performance loss. In this paper, we propose a sparse channel estimation method to exploit the sparse structure information in the CIR at each end user. Sparse channel estimation problem is formulated as compressed sensing (CS) using sparse decomposition theory and the estimation process is implemented by LASSO algorithm. Computer simulation results are given to confirm the superiority of proposed method over the LS-based channel estimation method.

cs.IT↗
arXiv2012

Improved Channel Estimation with Partial Sparse Constraint for AF Cooperative Communication Systems

Accurate channel state information (CSI) is necessary for coherent detection in amplify and forward (AF) broadband cooperative communication systems. Based on the assumption of ordinary sparse channel, efficient sparse channel estimation methods have been investigated in our previous works. However, when the cooperative channel exhibits partial sparse structure rather than ordinary sparsity, our previous method cannot take advantage of the prior information. In this paper, we propose an improved channel estimation method with partial sparse constraint on cooperative channel. At first, we formulate channel estimation as a compressive sensing problem and utilize sparse decomposition theory. Secondly, the cooperative channel is reconstructed by LASSO with partial sparse constraint. Finally, numerical simulations are carried out to confirm the superiority of proposed methods over ordinary sparse channel estimation methods.

cs.IT↗
arXiv2010

A Fast Compressive Channel Estimation with Modified Smoothed L0 Algorithm

Broadband wireless channel is a time dispersive and becomes strongly frequency selective. In most cases, the channel is composed of a few dominant coefficients and a large part of coefficients is approximately zero or zero. To exploit the sparsity of multi-path channel (MPC), there are various methods have been proposed. They are, namely, greedy algorithms, iterative algorithms, and convex program. The former two algorithms are easy to be implemented but not stable; on the other hand, the last method is stable but difficult to be implemented as practical channel estimation problems because of computational complexity. In this paper, we proposed a novel channel estimation strategy by using modified smoothed (MSL0) algorithm which combines stable and low complexity. Computer simulations confirm the effectiveness of the introduced algorithm comparisons with the existing methods. We also give

cs.IT↗
arXiv2010

Sparse Multipath Channel Estimation Using Compressive Sampling Matching Pursuit Algorithm

Wideband wireless channel is a time dispersive channel and becomes strongly frequency-selective. However, in most cases, the channel is composed of a few dominant taps and a large part of taps is approximately zero or zero. To exploit the sparsity of multi-path channel (MPC), two methods have been proposed. They are, namely, greedy algorithm and convex program. Greedy algorithm is easy to be implemented but not stable; on the other hand, the convex program method is stable but difficult to be implemented as practical channel estimation problems. In this paper, we introduce a novel channel estimation strategy using compressive sampling matching pursuit (CoSaMP) algorithm which was proposed in [1]. This algorithm will combine the greedy algorithm with the convex program method. The effectiveness of the proposed algorithm will be confirmed through comparisons with the existing methods.

cs.IT↗
arXiv2009

Epitaxial growth and magnetoelectric relaxor behavior in multiferroic 0.8Pb(Fe1/2Nb1/2)O3-0.2Pb(Mg1/2W1/2)O3 thin films

We present electric and magnetic properties of 0.8Pb(Fe1/2Nb1/2)O3-0.2Pb(Mg1/2W1/2)O3 films epitaxially grown on (001) SrTiO3 substrates using pulsed laser deposition. A narrow deposition window around 710 oC and 0.2 mbar has been identified to achieve epitaxial single-phase thin films. A typical Vogel-Fulcher relaxor-like dielectric and magnetic susceptibility dispersion is observed, suggesting magnetoelectric relaxor behavior in these films similar to the bulk. We determine a magnetic cluster freezing temperature of 36 K, while observing weak ferromagnetism via magnetic hysteresis loops up to 300 K.

cond-mat.mtrl-sci↗