Kai Zhong

Prof. Dr. Kai Zhong

Institutes of Physical Science and Information Technology, Anhui University, Anhui 230601, China
orcid

Research Focus & Metrics

Academic Career Length

6Years (2021-Present)

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

Representative Publications

View All Publications
31 May 2024 in IEEE Transactions on Reliability
Verified
Journal Article
Closed access
KZKai Zhong
Kai Zhong
ZDZhengping Ding
Zhengping Ding
HZHaifeng Zhang
Haifeng Zhang
Hongtian ChenHongtian Chen
Hongtian Chen
Enrico ZioEnrico Zio
Enrico Zio
Federated learning (FL)-based fault diagnosis is being widely developed. However, most of the existing FL methods may suffer from two drawbacks: 1) they are limited to a single diagnosis task, and this may be insufficient when comprehensive health status information is needed and 2) most of them work offline, thus neglecting the useful information contained in newly collected operation data. For this end, this article proposes a multitask federated incremental learning (multitask-FIL) framework. First of all, a multitask feature sharing network is established by assigning the extracted general features to different downstream tasks, so that the joint loss function is obtained for subsequent collaborative training. Then, Q-learning algorithm is used to select the incremental sequences for all the parties from real-time running data, which can facilitate the model performance by involving additional data information and preferred parties. After that, the incremental weight of each party is dynamically adjusted according to the loss depth and sample size in each round of communication, so that the effects of different parties can be quantified throughout the model iteration and aggregation process. Finally, experiments on three challenging cases are performed to show that the proposed method has strong multitask collaboration capability.

14 September 2022 in IEEE Transactions on Neural Networks and Learning Systems
Verified
Journal Article
Closed access
XDXiao Ding
Xiao Ding
HZHaifeng Zhang
Haifeng Zhang
CMChuang Ma
Chuang Ma
XZXingyi Zhang
Xingyi Zhang
KZKai Zhong
Kai Zhong
Recently, the problem of user identification across multiple social networks (UIAMSNs) has attracted considerable attention since it is a prerequisite for many downstream tasks and applications. Although substantial network feature-based approaches have been proposed to solve the UIAMSNs’ problem, the matching degree in most of the current works is given by experience, which lacks a solid theoretical basis. To alleviate the above predicament, we propose a user identification algorithm based on naive Bayes model (UI-NBM) within the network feature-based framework. First, a matching degree index is designed based on the naive Bayes model, which can accurately measure the contributions of different common matched node pairs (MNPs) to the connection probability of unmatched node pairs (UMNPs). Second, the matching degrees of all UMNPs are formulated as the product of matrices, giving rise to the great reduction of the time complexity and the compact expression; Finally, with the idea of recursion process, more UMNPs can be iteratively predicted even when only a small amount of prior information (i.e., a few number of MNPs) is known. The experimental results on the synthetic and real cross platforms demonstrate that the method outperforms the baseline methods within the feature-based framework.

2 August 2023 in Engineering Applications of Artificial Intelligence
Verified
Journal Article
Closed access
KZKai Zhong
Kai Zhong
JWJiayi Wang
Jiayi Wang
SXShuiqing Xu
Shuiqing Xu
CCChao Cheng
Chao Cheng
Hongtian ChenHongtian Chen
Hongtian Chen

23 November 2022 in IEEE Transactions on Systems, Man, and Cybernetics: Systems
Verified
Journal Article
Closed access
WHWen Hu
Wen Hu
XXXin Xia
Xin Xia
XDXiao Ding
Xiao Ding
XZXingyi Zhang
Xingyi Zhang
KZKai Zhong
Kai Zhong
HZHai-Feng Zhang
Hai-Feng Zhang
Existing methods regarding the influential nodes identification in complex networks usually assume that the structures of the networks are fully known. However, in many cases, knowing the full structure of one network is hard or impossible, and each participant can only obtain the partial structure of the networks. Therefore, each participant can be viewed as a private network (PN) of the original network, and they form multiple PNs. Then, one question arises: how to collaboratively identify influential nodes in multiple PNs while protecting the privacy of each PN. To this end, a secure multiparty computation ranking (SMPC-ranking) method is proposed to solve such an all-new problem based on the secure multiparty computation (SMPC) protocol, in which the SMPC protocol jointly computes a centrality function to measure the comprehensive importance of nodes or their ranking results by integrating each PN’s inputs without disclosing their inputs (i.e., the nodes’ importance). Experimental results in different networks demonstrate that, no matter what type of the centrality index is, the SMPC-ranking method can better identify the influential nodes in networks than that of the baseline method only using the structural information of one PN. What is more, our SMPC-ranking method does not violate the privacy of each PN. Therefore, the proposed method provides new insights on collaboratively identifying influential nodes in multiple PNs and has potential applications in many fields.

7 March 2023 in IEEE/ASME Transactions on Mechatronics
Verified
Journal Article
Closed access
KZKai Zhong
Kai Zhong
BHBing Han
Bing Han
MHMin Han
Min Han
Hongtian ChenHongtian Chen
Hongtian Chen
Industrial sensor signals are essentially non-Euclidean graph structures due to the interplay between process variables; thus, graph convolutional networks (GCNs) have been widely studied and applied. However, most of the existing GCN-based methods may suffer from two drawbacks: 1) it is difficult to characterize multiple interactions among nodes and 2) the input graph constructed from the original data may contain errors and missing edges, which will degenerate the fault diagnosis performance. To address the abovementioned issues, this article designs a hierarchical GCN with latent structure learning for industrial fault diagnosis, which can organize hierarchical networks to collaboratively improve the quality of latent graph structure, and enhanced diagnostic performance can be guaranteed. First, a high-quality updated graph is formed by incorporating the original graph with the new graph in the graph constructing layer, which can not only eliminate the adverse effects of noise and outliers but also characterize the multiple interactions among nodes. Then, the updated graph is fed into the multilayer GCN layer for better feature learning and enhances the node representation through intra- and inter-layer convolutional operations simultaneously. After that, the produced node embeddings are used to guide the latent structure learning process for optimal graph. Finally, the proposed method is verified in both the simulated and real industrial processes. The experimental results demonstrate that the new approach has better fault diagnosis accuracy and practicability than state-of-the-art methods.

22 May 2025 in IEEE Transactions on Power Electronics
Verified
Journal Article
Closed access
KZKai Zhong
Kai Zhong
ZLZhihao Liu
Zhihao Liu
Jiaqiang TianJiaqiang Tian
Jiaqiang Tian
CFChao Fan
Chao Fan
Mince LiMince Li
Mince Li
Peng WangPeng Wang
Peng Wang
The state of health (SOH) and remaining useful life (RUL) prediction of lithium-ion batteries are critical for the safe operation and maintenance optimization of electrical systems. However, existing methods face challenges in generalization ability and prediction accuracy, including the limitations of single-task learning for joint SOH and RUL prediction, sensitivity to complex operating conditions, and the impact of cross-domain data distribution differences. To address these challenges, this article proposes a dual-representation transferable knowledge sharing network to predict SOH and RUL for batteries. The proposed algorithm is based on the framework of multi-task learning and transfer learning. By extracting and fusing dual-representation data features, the framework effectively captures the latent patterns of battery aging characteristics. It employs parameter and network structure sharing mechanisms to enhance the adaptability and generalization of cross-domain features. The proposed model achieves information collaboration between SOH and RUL tasks through a shared network, effectively reducing computational costs and improving prediction accuracy. Experimental results demonstrate that the proposed framework exhibits excellent robustness and transferability in complex degradation scenarios, significantly outperforming baseline models.