Jinya Zhang

Jinya Zhang

Auto-generated profile, based on publicly available information

Jinya Zhang

School of Information Science and Engineering, Southeast University, Nanjing, China

Research Focus & Metrics

Academic Career Length

24Years (2003-Present)

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

Representative Publications

View All Publications
8 December 2003 in Physical Review B
Verified
Journal Article
Closed access
PDF file
CJChao Jing
Chao Jing
SCS. X. Cao
S. X. Cao
JZJ. C. Zhang
J. C. Zhang
The dependence of magnetic properties on lattice constants in both body-centered-cubic (bcc) and face-centered-cubic (fcc) FexMn1−x (x=0,0.25,0.5,0.75,1) has been studied systematically using first-principles linearized augmented plane-wave (LAPW) calculations. The results indicate that the magnetic properties of bcc FexMn1−x alloys are ferromagnetism within a fairly large range of lattice constants, and the atomic moment increases with an increase of the lattice constants. However, for fcc FexMn1−x alloys, the magnetic properties change from paramagnetism to antiferromagnetism (or ferrimagnetism) and finally to ferromagnetism with an increase of the lattice constants. The calculated magnetic moment is consistent with previous experiments for bcc FexMn1−x alloys at the normal lattice constants, while for fcc FexMn1−x alloys on the whole the magnetic properties are also in agreement with the experimental results. The present results may facilitate a better understanding of the correlation between the structures and the corresponding magnetic properties.

30 August 2024 in IEEE Wireless Communications Letters
Verified
Journal Article
Closed access
JZJinya Zhang
Jinya Zhang
JGJiajia Guo
Jiajia Guo
CWChao-Kai Wen
Chao-Kai Wen
SJShi Jin
Shi Jin
The utilization of a Deep Autoencoder (DAE) in End-to-End (E2E) communication systems has been recognized for its potential to achieve global optimality, especially in scenarios involving multiple antennas. Although previous studies have leveraged channel state information (CSI) at the transmitter or receiver to augment DAE performance, a comprehensive examination of the CSI acquisition process remains limited. Addressing this gap, this paper introduces a task-driven approach to CSI acquisition for E2E communications with multiple antennas. The cornerstone of this approach is the selective acquisition of channel information pertinent to the design of the transceiver by integrating the CSI acquisition process with DAE, aiming to curtail CSI acquisition overhead. Implemented through AI-based pilot symbols generation and received pilot signals feedback, this method avoids explicit channel estimation and CSI reconstruction at the transceiver, facilitating a joint design paradigm. Our simulation results underline that the proposed framework substantially outperforms traditional maximum ratio transmission (MRT) precoding systems and other advanced DAE models in terms of both reducing channel acquisition overhead and improving overall system performance.

23 September 2025 in IEEE Transactions on Wireless Communications
Verified
Journal Article
Closed access
JZJinya Zhang
Jinya Zhang
JGJiajia Guo
Jiajia Guo
XLXiangyi Li
Xiangyi Li
CWChao-Kai Wen
Chao-Kai Wen
Xin GengXin Geng
Xin Geng
SJShi Jin
Shi Jin
Deep learning (DL) has introduced a new paradigm in multiple-input multiple-output (MIMO) detection, balancing performance and complexity. However, the practical deployment of DL-based detectors is hindered by poor generalization, high training costs, and inflexibility in varying-scale models construction. To address these challenges, this paper presents a novel knowledge transfer technique, termed learngene, for the design of a DL-based MIMO detector and proposes an efficient deployment framework. The proposed detector, SDNet, leverages zero-forcing detection outputs and least squares-estimated channel state information (CSI) as inputs. It is further optimized through a collective-individual paradigm to enhance knowledge transfer. In this paradigm, learngene, a reusable neural network (NN) segment, encapsulates detection meta-knowledge acquired from large-scale collective models trained by manufacturers. This segment can then be distributed to device-specific teams. By integrating learngene into different lightweight individual models, detection meta-knowledge is efficiently transferred across heterogeneous NNs, enabling adaptation to diverse devices and scenarios. Simulation results demonstrate that the proposed scheme enhances performance, enables rapid adaptation, and ensures high scalability, with transferred parameters comprising only 10.8% of the total model size.