Jun Wu

Jun Wu

Auto-generated profile, based on publicly available information

Jun Wu

Fudan University

Research Focus & Metrics

Academic Career Length

12Years (2015-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
10 July 2023 in IEEE Transactions on Information Forensics and Security
Verified
Journal Article
Closed access
KWKang Wei
Kang Wei
JLJun Li
Jun Li
CMChuan Ma
Chuan Ma
MDMing Ding
Ming Ding
Wen ChenWen Chen
Wen Chen
H. Vincent PoorH. Vincent Poor
H. Vincent Poor
Personalized federated learning (PFL), as a novel federated learning (FL) paradigm, is capable of generating personalized models for heterogenous clients. Combined with with a meta-learning mechanism, PFL can further improve the convergence performance with few-shot training. However, meta-learning based PFL has two stages of gradient descent in each local training round, therefore posing a more serious challenge in information leakage. In this paper, we propose a differential privacy (DP) based PFL (DP-PFL) framework and analyze its convergence performance. Specifically, we first design a privacy budget allocation scheme for inner and outer update stages based on the Rényi DP composition theory. Then, we develop two convergence bounds for the proposed DP-PFL framework under convex and non-convex loss function assumptions, respectively. Our developed convergence bounds reveal that 1) there is an optimal size of the DP-PFL model that can achieve the best convergence performance for a given privacy level, and 2) there is an optimal tradeoff among the number of communication rounds, convergence performance and privacy budget. Evaluations on various real-life datasets demonstrate that our theoretical results are consistent with experimental results. The derived theoretical results can guide the design of various DP-PFL algorithms with configurable tradeoff requirements on the convergence performance and privacy levels.

27 June 2018 in IEEE Internet of Things Journal
Verified
Journal Article
Closed access
QZQingqing Zhang
Qingqing Zhang
WFWen Fang
Wen Fang
QLQingwen Liu
Qingwen Liu
JWJun Wu
Jun Wu
PXPengfei Xia
Pengfei Xia
LYLiuqing Yang
Liuqing Yang
Wireless power transfer (WPT) is a promising solution to provide convenient and perpetual energy supplies to electronics. Traditional WPT technologies face the challenge of providing Watt-level power over meter-level distance for Internet of Things (IoT) and mobile devices, such as sensors, controllers, smart-phones, laptops, etc.. Distributed laser charging (DLC), a new WPT alternative, has the potential to solve these problems and enable WPT with the similar experience as WiFi communications. In this paper, we present a multi-module DLC system model, in order to illustrate its physical fundamentals and mathematical formula. This analytical modeling enables the evaluation of power conversion or transmission for each individual module, considering the impacts of laser wavelength, transmission attenuation and photovoltaic-cell (PV-cell) temperature. Based on the linear approximation of electricity-to-laser and laser-to-electricity power conversion validated by measurement and simulation, we derive the maximum power transmission efficiency in closed-form. Thus, we demonstrate the variation of the maximum power transmission efficiency depending on the supply power at the transmitter, laser wavelength, transmission distance, and PV-cell temperature. Similar to the maximization of information transmission capacity in wireless information transfer (WIT), the maximization of the power transmission efficiency is equally important in WPT. Therefore, this work not only provides the insight of DLC in theory, but also offers the guideline of DLC system design in practice.

20 September 2019 in IEEE Wireless Communications Letters
Verified
Journal Article
Closed access
DHDie Hu
Die Hu
YZYonghao Zhang
Yonghao Zhang
LHLianghua He
Lianghua He
JWJun Wu
Jun Wu
This letter proposes a low-complexity deep-learning-based direction-of-arrival (DOA) estimation method for a hybrid massive multiple-input multiple-output (MIMO) system with a uniform circular array at the base station. In the proposed method, we first input the received signal vector into some small deep feedforward networks that are trained offline. Based on the outputs of the networks, we then generate a set of candidate angles. By selecting the optimal one from all candidate angles, we finally obtain the DOA estimation. Simulation results demonstrate that, compared with the conventional maximum likelihood (ML) method, the proposed DOA estimation method can achieve similar or even better performance with much less complexity.

30 March 2021 in IEEE Transactions on Wireless Communications
Verified
Journal Article
Closed access
ZXZhe Xing
Zhe Xing
RWRui Wang
Rui Wang
JWJun Wu
Jun Wu
ELErwu Liu
Erwu Liu
Intelligent reflecting surface (IRS) is envisioned as a promising hardware solution to hardware cost and energy consumption in the fifth-generation (5G) mobile communication network. It exhibits great advantages in enhancing data transmission, but may suffer from performance degradation caused by inherent hardware impairment (HWI). For analysing the achievable rate (ACR) and optimizing the phase shifts in the IRS-aided wireless communication system with HWI, we consider that the HWI appears at both the IRS and the signal transceivers. On this foundation, first, we derive the closed-form expression of the average ACR and the IRS utility. Then, we formulate optimization problems to optimize the IRS phase shifts by maximizing the signal-to-noise ratio (SNR) at the receiver side, and obtain the solution by transforming non-convex problems into semidefinite programming (SDP) problems. Subsequently, we compare the IRS with the conventional decode-and-forward (DF) relay in terms of the ACR and the utility. Finally, we carry out simulations to verify the theoretical analysis, and evaluate the impact of the channel estimation errors and residual phase noises on the optimization performance. Our results reveal that the HWI reduces the ACR and the IRS utility, and begets more serious performance degradation with more reflecting elements. Although the HWI has an impact on the IRS, it still leaves opportunities for the IRS to surpass the conventional DF relay, when the number of reflecting elements is large enough or the transmitting power is sufficiently high.

6 April 2021 in IEEE Transactions on Image Processing
Verified
Journal Article
Closed access
Cairong ZhaoCairong Zhao
Cairong Zhao
XLXinbi Lv
Xinbi Lv
SDShuguang Dou
Shuguang Dou
SZShanshan Zhang
Shanshan Zhang
JWJun Wu
Jun Wu
LWLiang Wang
Liang Wang
Person re-identification (re-id) suffers from the significant challenge of occlusion, where an image contains occlusions and less discriminative pedestrian information. However, certain work consistently attempts to design complex modules to capture implicit information (including human pose landmarks, mask maps, and spatial information). The network, consequently, focuses on discriminative features learning on human non-occluded body regions and realizes effective matching under spatial misalignment. Few studies have focused on data augmentation, given that existing single-based data augmentation methods bring limited performance improvement. To address the occlusion problem, we propose a novel Incremental Generative Occlusion Adversarial Suppression (IGOAS) network. It consists of 1) an incremental generative occlusion block, generating easy-to-hard occlusion data, that makes the network more robust to occlusion by gradually learning harder occlusion instead of hardest occlusion directly. And 2) a global-adversarial suppression (G&A) framework with a global branch and an adversarial suppression branch. The global branch extracts steady global features of the images. The adversarial suppression branch, embedded with two occlusion suppression module, minimizes the generated occlusion’s response and strengthens attentive feature representation on human non-occluded body regions. Finally, we get a more discriminative pedestrian feature descriptor by concatenating two branches’ features, which is robust to the occlusion problem. The experiments on the occluded dataset show the competitive performance of IGOAS. On Occluded-DukeMTMC, it achieves 60.1% Rank-1 accuracy and 49.4% mAP.

3 November 2020 in International Journal of Computer Vision
Verified
Journal Article
Closed access
PDF file
XCXin Chen
Xin Chen
LXLingxi Xie
Lingxi Xie
JWJun Wu
Jun Wu
QTQi Tian
Qi Tian