Changwei Zhang

Dr. Changwei Zhang

orcid

Research Focus & Metrics

Representative Publications

View All Publications
22 September 2022 in IEEE Transactions on Wireless Communications
Verified
Journal Article
Closed access
CZChangwei Zhang
Changwei Zhang
XSXinghua Sun
Xinghua Sun
Wenchao XiaWenchao Xia
Wenchao Xia
JZJun Zhang
Jun Zhang
HZHongbo Zhu
Hongbo Zhu
Xianbin WangXianbin Wang
Xianbin Wang
With the rapid development of 5G, massive machine-type communication is expected to experience significant growth, leading to severe random access collisions. To address this issue, we first adopt a deep neural network to detect random access collisions by learning the features of the received signals. Based on the collision-detection results, we propose a double-contention random access (DCRA) scheme, with which the base station can schedule one more contention process for devices experiencing collisions. To fully harness the collision-resolution capability of the proposed DCRA scheme, we further analyze its performance and illustrate how to tune the backoff parameters to optimize the network throughput. It is revealed that the maximum throughput of the DCRA scheme depends on the number of random access preambles and the collision recognition accuracy. The corresponding optimal backoff parameters are then obtained, which greatly facilitates implementations in practice. Simulation results show that with a high collision recognition accuracy, the proposed scheme can achieve significant throughput improvement.

27 August 2020 in IEEE Transactions on Vehicular Technology
Verified
Journal Article
Closed access
CZChangwei Zhang
Changwei Zhang
XSXinghua Sun
Xinghua Sun
JZJun Zhang
Jun Zhang
Xianbin WangXianbin Wang
Xianbin Wang
SJShi Jin
Shi Jin
HZHongbo Zhu
Hongbo Zhu
Machine-to-machine (M2M) communications, which aims to enable autonomous communication between machine devices with minimum human intervention, have experienced remarkable growth over the past few years. In massive M2M communications, random access (RA) collision is one of critical problems at the network entry stage. To control the collision rate during the network access stage, most existing RA solutions for M2M communications need the base station (BS) to determine the optimal backoff parameters, which are hard to obtain and also would incur substantial signaling overhead, especially in dense scenarios. To address this issue, a distributive approach is highly desirable. In this paper, a new method for distributive estimation of optimal backoff parameters is proposed. Specifically, each MTD only needs to observe their own successful and total transmissions of access requests during an estimation interval, after which each MTD can obtain an estimation of the optimal backoff parameters according to an explicit expression of the observed statistics. It is found that the estimation interval determines the performance of the proposed scheme. When the number of MTDs in the network does not change or varies slowly, the network throughput can be maximized if the estimation interval is chosen properly. Compared to existing centralized methods, the proposed method can achieve similar performance with lower signaling overhead. The proposed scheme can be applied for both homogeneous and heterogeneous scenarios and provide a practical access design method for M2M communications.

26 October 2022 in IEEE Wireless Communications Letters
Verified
Journal Article
Closed access
MZMeng Zhou
Meng Zhou
JLJinwen Li
Jinwen Li
MXMangang Xie
Mangang Xie
CZChangwei Zhang
Changwei Zhang
JYJiantao Yuan
Jiantao Yuan
This letter proposes a device-to-device (D2D) underlaid multigroup multicast cell-free massive multi-input multi-output (MIMO) system with low-resolution digital-to-analog converters (DACs) at distributed access points (APs) and multiple antennas at both the APs and users. The closed-form rate expressions for cell-free users (CFUEs) and D2D users (DUEs) are derived using imperfect channel state information (CSI). Finally, simulation results for the achievable average sum rate shed light on the impact of key parameters and highlight the significance of the collaborative deployment of D2D and multigroup multicast cell-free massive MIMO.

12 May 2025 in Journal of Medical Internet Research
Verified
Journal Article
Open access
RHRuochen Huang
Ruochen Huang
HWHonghan Wu
Honghan Wu
YYYuhan Yuan
Yuhan Yuan
YXYifan Xu
Yifan Xu
HQHao Qian
Hao Qian
YLYun Liu
Yun Liu
Background: Synthetic electronic health records (EHRs) generated by large language models (LLMs) offer potential for clinical education and model training while addressing privacy concerns. However, performance variations and demographic biases in these models remain underexplored, posing risks to equitable health care. Objective: This study aimed to systematically assess the performance of various LLMs in generating synthetic EHRs and to critically evaluate the presence of gender and racial biases in the generated outputs. We focused on assessing the completeness and representativeness of these EHRs across 20 diseases with varying demographic prevalence. Methods: A framework was developed to generate 140,000 synthetic EHRs using 10 standardized prompts across 7 LLMs. The electronic health record performance score (EPS) was introduced to quantify completeness, while the statistical parity difference (SPD) was proposed to assess the degree and direction of demographic bias. Chi-square tests were used to evaluate the presence of bias across demographic groups. Results: Larger models exhibited superior performance but heightened biases. The Yi-34B achieved the highest EPS (96.8), while smaller models (Qwen-1.8B: EPS=63.35) underperformed. Sex polarization emerged: female-dominated diseases (eg, multiple sclerosis) saw amplified female representation in outputs (Qwen-14B: 973/1000, 97.3% female vs 564,424/744,778, 75.78% real; SPD=+21.50%), while balanced diseases and male-dominated diseases skewed the male group (eg, hypertension Llama 2-13 B: 957/1000, 95.7% male vs 79,540,040/152,466,669, 52.17% real; SPD=+43.50%). Racial bias patterns revealed that some models overestimated the representation of White (eg, Yi-6B: mean SPD +14.40%, SD 16.22%) or Black groups (eg, Yi-34B: mean SPD +14.90%, SD 27.16%), while most models systematically underestimated the representation of Hispanic (average SPD across 7 models is –11.93%, SD 8.36%) and Asian groups (average SPD across 7 models is –0.77%, SD 11.99%). Conclusions: Larger models, such as Yi-34B, Qwen-14B, and Llama 2 to 13 B, showed improved performance in generating more comprehensive EHRs, as reflected in higher EPS values. However, this increased performance was accompanied by a notable escalation in both gender and racial biases, highlighting a performance-bias trade-off. The study identified 4 key findings as follows: (1) as model size increased, EHR generation improved, but demographic biases also became more pronounced; (2) biases were observed across all models, not just the larger ones; (3) gender bias closely aligned with real-world disease prevalence, while racial bias was evident in only a subset of diseases; and (4) racial biases varied, with some diseases showing overrepresentation of White or Black populations and underrepresentation of Hispanic and Asian groups. These findings underline the need for effective bias mitigation strategies and the development of benchmarks to ensure fairness in artificial intelligence applications for health care.

25 July 2025 in IEEE Internet of Things Journal
Verified
Journal Article
Closed access
RHRuochen Huang
Ruochen Huang
CQChen Qiu
Chen Qiu
MCMingkai Chen
Mingkai Chen
CZChangwei Zhang
Changwei Zhang
HZHongbo Zhu
Hongbo Zhu
In view of super-large scale access and dynamic connectivity requirements in 6G, the number of users and data is increasing exponentially, which makes it difficult to achieve sustainable development of communications. Meanwhile, with the development of cross-modal processing, semantic information is highly considered accurate and intelligent, which is expected to bring new ideas for 6G. Therefore, in this study, we propose a novel framework of cross-modal semantic communications to improve the efficiency of the multi-modal processing, especially the tactile modal. Meanwhile, we have redesigned the significant aspects of artificial intelligence (AI), such as encoding, transmission, and processing. As federated learning (FL) inherently supports multiple privacy-preserving and security measures, we also introduce FL to assist AI training in cross-modal semantic communications, including the expansion of multi-modal data, cross-modal semantic extraction, comprehensive decision-making, and privacy protection. First, in encoding aspects, we propose a hybrid coding for haptic signal coding by deep learning (DL) according to the semantic association between material identification and tactile code optimization. Second, in transmission aspects, we propose a modal-aware resource allocation for the fairness optimization between transmission requirements and network resources with deep reinforcement learning (DRL). Third, in signal processing, low-rank signal reconstruction and immersive quality of experience (QoE) evaluation by DL and machine learning (ML) are provided to improve and quantify users’ experience. In addition, the computational experiments of those key technologies are shown individually and the specification of their probability is discussed separately. Finally, future research topics related to the above issues are suggested.

29 May 2023 in IEEE Communications Letters
Verified
Journal Article
Closed access
CZChangwei Zhang
Changwei Zhang
XSXinghua Sun
Xinghua Sun
Wenchao XiaWenchao Xia
Wenchao Xia
RHRuochen Huang
Ruochen Huang
MZMeng Zhou
Meng Zhou
HZHongbo Zhu
Hongbo Zhu
Machine learning-based random access schemes have gained significant attention in recent years. However, optimizing the access delay of such schemes remains a challenge. In this letter, we focus on analyzing and optimizing the mean access delay of the deep neural network-based double-contention random access (DCRA) scheme, proposed in our earlier work. Specifically, by leveraging the characterization of the state transition of each access request, the mean sojourn time of each state can be derived explicitly, based on which the mean access delay of the DCRA scheme is further obtained. To optimize the mean access delay, we propose optimizing the backoff parameters of every MTDs. Simulation results demonstrate that the mean access delay of the DCRA scheme can be significantly reduced compared to existing schemes, by appropriately selecting the backoff parameters of every MTDs.