Zhiyong Chen
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Zhiyong Chen has a total of 188 co-authors. Here are the top co-authors by number of shared publications:
- Meixia Tao
Shanghai Jiao Tong University
35 shared publications - Bin Xia
Shanghai Jiao Tong University
31 shared publications - Wenjun Zhang
Shanghai Jiao Tong University
16 shared publications - Hui Liu
Department of Electrical and Computer Engineering University of Washington
12 shared publications - Yin Xu
Shanghai Jiao Tong University
12 shared publications - Dazhi He
Shanghai Jiao Tong University
11 shared publications
Latest Publications
- 1 December 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Mobile Computing
- Vol. 25  (12) , 24369-24385
PODD: Path-Optimized Denoising Diffusion Models for Wireless Semantic Communications
- 1 December 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Mobile Computing
- Vol. 25  (12) , 24369-24385
PODD: Path-Optimized Denoising Diffusion Models for Wireless Semantic Communications
- 1 October 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Wireless Communications
- Vol. 33  (5) , 62-72
Generative Communications: Overview, Technologies, and Trends
- 1 October 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Wireless Communications
- Vol. 33  (5) , 62-72
Generative Communications: Overview, Technologies, and Trends
- 8 September 2026
- Published by Springer Nature in Science China Information Sciences
- Vol. 69  (10) , 200305
E3-Agent: an executable and evolving agent for resource management of edge generative inference
- 8 September 2026
- Published by Springer Nature in Science China Information Sciences
- Vol. 69  (10) , 200305
E3-Agent: an executable and evolving agent for resource management of edge generative inference
- Open Access
- 12 May 2026
- Published by MDPI AG in Preprints.org
A Comparative Survey of Inference Acceleration for DLLMs against AR-LLMs: No Free Lunch
Autoregressive large language models (AR-LLMs) have achieved remarkable success, but their inherently sequential decoding process remains a fundamental bottleneck for efficient inference. Diffusion large language models (DLLMs), with bidirectional modeling and parallel token generation, offer a promising alternative to break this token-by-token limitation. Yet despite rapid progress, the practical inference efficiency of current DLLMs remains unclear. From a verification perspective, this survey establishes a systematic taxonomy of existing acceleration methods, benchmarks representative techniques under a unified experimental setting, and further evaluates strong strategy combinations to quantify the gap between mainstream DLLM inference methods and state-of-the-art AR baselines. Specially, the overall analysis highlights that the parallel decoding efficiency of DLLMs still remains a significant lag compared to the decoding efficiency of AR-LLMs under inference acceleration. We provide an in-depth experimental analysis about the underlying trade-offs among generation quality, latency, and system compatibility, and build up a standard evaluation bench open to the community. Remaining bottlenecks are also summarized, together with future directions for more practical and competitive DLLM inference. Code is available at \url{https://github.com/haoyun-jiang/DLLM-AccelEval}.
- 12 May 2026
- Published by MDPI AG in Preprints.org
Open AccessA Comparative Survey of Inference Acceleration for DLLMs against AR-LLMs: No Free Lunch
Autoregressive large language models (AR-LLMs) have achieved remarkable success, but their inherently sequential decoding process remains a fundamental bottleneck for efficient inference. Diffusion large language models (DLLMs), with bidirectional modeling and parallel token generation, offer a promising alternative to break this token-by-token limitation. Yet despite rapid progress, the practical inference efficiency of current DLLMs remains unclear. From a verification perspective, this survey establishes a systematic taxonomy of existing acceleration methods, benchmarks representative techniques under a unified experimental setting, and further evaluates strong strategy combinations to quantify the gap between mainstream DLLM inference methods and state-of-the-art AR baselines. Specially, the overall analysis highlights that the parallel decoding efficiency of DLLMs still remains a significant lag compared to the decoding efficiency of AR-LLMs under inference acceleration. We provide an in-depth experimental analysis about the underlying trade-offs among generation quality, latency, and system compatibility, and build up a standard evaluation bench open to the community. Remaining bottlenecks are also summarized, together with future directions for more practical and competitive DLLM inference. Code is available at \url{https://github.com/haoyun-jiang/DLLM-AccelEval}.
- 23 March 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Vehicular Technology
- Vol. PP  (00189545) , 1-5
Stable-MoE: Lyapunov-based Token Routing for Distributed Mixture-of-Experts Training over Edge Networks
The sparse activation mechanism of mixture of experts (MoE) model empowers edge intelligence with enhanced training efficiency and reduced computational resource consumption. However, traditional token routing in distributed MoE training faces significant challenges in resource-constrained edge net works characterized by heterogeneous computing capabilities and stochastic token arrivals, which inevitably suffer from workload backlog, resource inefficiency, and performance degradation. To address this issue, we propose a novel Lyapunov-based token routing framework for distributed MoE training over resource heterogeneous edge networks, termed Stable-MoE. Specifically, we formulate a stochastic optimization problem to maximize both system throughput and gating consistency via optimizing the token routing strategy and computational resource allocation, while ensuring long-term stability of both token and energy queues at the edge devices. Using the Lyapunov optimization, we transform the intractable long-term optimization problem into tractable per-slot subproblems by enabling online decision-making of token routing and computation frequency utilization without the knowledge of future system states. Experimental results on the SVHN and CIFAR-100 datasets demonstrate that Stable-MoE outperforms the baselines with at least 40% and 5% gains in system throughput and test accuracy, respectively.
- 23 March 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Vehicular Technology
- Vol. PP  (00189545) , 1-5
Stable-MoE: Lyapunov-based Token Routing for Distributed Mixture-of-Experts Training over Edge Networks
The sparse activation mechanism of mixture of experts (MoE) model empowers edge intelligence with enhanced training efficiency and reduced computational resource consumption. However, traditional token routing in distributed MoE training faces significant challenges in resource-constrained edge net works characterized by heterogeneous computing capabilities and stochastic token arrivals, which inevitably suffer from workload backlog, resource inefficiency, and performance degradation. To address this issue, we propose a novel Lyapunov-based token routing framework for distributed MoE training over resource heterogeneous edge networks, termed Stable-MoE. Specifically, we formulate a stochastic optimization problem to maximize both system throughput and gating consistency via optimizing the token routing strategy and computational resource allocation, while ensuring long-term stability of both token and energy queues at the edge devices. Using the Lyapunov optimization, we transform the intractable long-term optimization problem into tractable per-slot subproblems by enabling online decision-making of token routing and computation frequency utilization without the knowledge of future system states. Experimental results on the SVHN and CIFAR-100 datasets demonstrate that Stable-MoE outperforms the baselines with at least 40% and 5% gains in system throughput and test accuracy, respectively.
- 19 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Broadcasting
- Vol. 72  (1) , 96-105
A Frequency-Domain Feature-Based Self-Interference Cancellation Method for In-Band Full-Duplex Communications in Inter-Tower Communication Networks
As a hot research topic for next-generation broadcasting systems, enhancements to In-band Distribution Links (IDL) and the Inter-Tower Communications Network (ITCN) have attracted significant attention. With its significant advantages in spectrum efficiency and inter-tower wireless connectivity, the In-Band Full-Duplex (IBFD) technique has become a much-needed solution for Single-Frequency Network (SFN) signal backhaul, network synchronization and control, and emergency alert information distribution. However, the self-interference resulting from the local transmitter signal loop-back is the primary bottleneck impacting the performance of the IBFD technique, and the non-linear distortion caused by the high-power amplifier further complicates the cancellation of the self-interference. This paper proposes a full-duplex self-interference cancellation (SIC) method. Based on radio frequency (RF) reference signals and short-term learning via phase search and tracking, combined with prior knowledge of Advanced Television Standards Committee (ATSC) 3.0 signals, this method achieves good SIC performance with low complexity. Simulation results show that the proposed method achieves a good SIC gain of up to 46 dB in digital domain cancellation. With analog pre-cancellation and other co-channel interference mitigation techniques, 60+ dB total SIC cancellation can be achieved for ITCN applications. Meanwhile, the simulation results also show that this method can effectively address phase jitter and signal distortion caused by the distributed parameters of hardware devices and environmental variations.
- 19 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Broadcasting
- Vol. 72  (1) , 96-105
A Frequency-Domain Feature-Based Self-Interference Cancellation Method for In-Band Full-Duplex Communications in Inter-Tower Communication Networks
As a hot research topic for next-generation broadcasting systems, enhancements to In-band Distribution Links (IDL) and the Inter-Tower Communications Network (ITCN) have attracted significant attention. With its significant advantages in spectrum efficiency and inter-tower wireless connectivity, the In-Band Full-Duplex (IBFD) technique has become a much-needed solution for Single-Frequency Network (SFN) signal backhaul, network synchronization and control, and emergency alert information distribution. However, the self-interference resulting from the local transmitter signal loop-back is the primary bottleneck impacting the performance of the IBFD technique, and the non-linear distortion caused by the high-power amplifier further complicates the cancellation of the self-interference. This paper proposes a full-duplex self-interference cancellation (SIC) method. Based on radio frequency (RF) reference signals and short-term learning via phase search and tracking, combined with prior knowledge of Advanced Television Standards Committee (ATSC) 3.0 signals, this method achieves good SIC performance with low complexity. Simulation results show that the proposed method achieves a good SIC gain of up to 46 dB in digital domain cancellation. With analog pre-cancellation and other co-channel interference mitigation techniques, 60+ dB total SIC cancellation can be achieved for ITCN applications. Meanwhile, the simulation results also show that this method can effectively address phase jitter and signal distortion caused by the distributed parameters of hardware devices and environmental variations.
- Open Access
Joint Source-Channel-Generation Coding: From Distortion-oriented Reconstruction to Semantic-consistent Generation
Conventional communication systems, including both separation-based coding and AI-driven joint source-channel coding (JSCC), are largely guided by Shannon's rate-distortion theory. However, relying on generic distortion metrics fails to capture complex human visual perception, often resulting in blurred or unrealistic reconstructions. In this paper, we propose Joint Source-Channel-Generation Coding (JSCGC), a novel paradigm that shifts the focus from deterministic reconstruction to probabilistic generation. JSCGC leverages a generative model at the receiver as a generator rather than a conventional decoder to parameterize the data distribution, enabling direct maximization of mutual information under channel constraints while controlling stochastic sampling to produce outputs residing on the authentic data manifold with high fidelity. We further derive a theoretical lower bound on the maximum semantic inconsistency with given transmitted mutual information, elucidating the fundamental limits of communication in controlling the generative process. Extensive experiments on image transmission demonstrate that JSCGC substantially improves perceptual quality and semantic fidelity, significantly outperforming conventional distortion-oriented JSCC methods.
Open AccessJoint Source-Channel-Generation Coding: From Distortion-oriented Reconstruction to Semantic-consistent Generation
Conventional communication systems, including both separation-based coding and AI-driven joint source-channel coding (JSCC), are largely guided by Shannon's rate-distortion theory. However, relying on generic distortion metrics fails to capture complex human visual perception, often resulting in blurred or unrealistic reconstructions. In this paper, we propose Joint Source-Channel-Generation Coding (JSCGC), a novel paradigm that shifts the focus from deterministic reconstruction to probabilistic generation. JSCGC leverages a generative model at the receiver as a generator rather than a conventional decoder to parameterize the data distribution, enabling direct maximization of mutual information under channel constraints while controlling stochastic sampling to produce outputs residing on the authentic data manifold with high fidelity. We further derive a theoretical lower bound on the maximum semantic inconsistency with given transmitted mutual information, elucidating the fundamental limits of communication in controlling the generative process. Extensive experiments on image transmission demonstrate that JSCGC substantially improves perceptual quality and semantic fidelity, significantly outperforming conventional distortion-oriented JSCC methods.
- 1 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Communications
- Vol. 74, 12654-12670
Spatial Semantic Communication: When Semantic Transmission Meets Index Modulation
- 1 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Communications
- Vol. 74, 12654-12670
Spatial Semantic Communication: When Semantic Transmission Meets Index Modulation
- Top Cited Papers
- 1 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Wireless Communications
- Vol. 25, 559-572
WDMoE: Wireless Distributed Mixture of Experts for Large Language Models
- 1 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Wireless Communications
- Vol. 25, 559-572
Top Cited PapersWDMoE: Wireless Distributed Mixture of Experts for Large Language Models
- 22 December 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Wireless Communications
- Vol. 25  (15361276) , 9264-9279
MambaJSCC: Adaptive Deep Joint Source–Channel Coding With Generalized State Space Model
Lightweight and efficient neural network models for deep joint source-channel coding (JSCC) are crucial for semantic communications. In this paper, we propose a novel JSCC architecture, named MambaJSCC, that achieves great performance with low computational and parameter overhead. MambaJSCC utilizes the visual state space model with channel adaptation (VSSM-CA) blocks as its backbone for transmitting images over wireless channels, where the VSSM-CA primarily consists of the generalized state space models (GSSM) and the zero-parameter, zero-computational channel adaptation method (CSI-ReST). We design the GSSM module, leveraging reversible matrix transformations to express generalized scan expanding operations, and theoretically prove that two GSSM modules can effectively capture global information. We discover that GSSM inherently possesses the ability to adapt to channels, a form of endogenous intelligence. Based on this, we design the CSI-ReST method, which injects channel state information (CSI) into the initial state of GSSM to utilize its native response, and into the residual state to mitigate CSI forgetting, enabling effective channel adaptation without introducing additional computational and parameter overhead. Experimental results on different devices, including IoT device JETSON AGX ORIN, show that MambaJSCC not only outperforms existing JSCC methods (e.g., SwinJSCC) across various scenarios but also significantly reduces parameter size, computational overhead, and inference delay. We have released our code and pre-trained models at https://github.com/Wireless3C-SJTU/MambaJSCC, allowing full reproduction of our results.
- 22 December 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Wireless Communications
- Vol. 25  (15361276) , 9264-9279
MambaJSCC: Adaptive Deep Joint Source–Channel Coding With Generalized State Space Model
Lightweight and efficient neural network models for deep joint source-channel coding (JSCC) are crucial for semantic communications. In this paper, we propose a novel JSCC architecture, named MambaJSCC, that achieves great performance with low computational and parameter overhead. MambaJSCC utilizes the visual state space model with channel adaptation (VSSM-CA) blocks as its backbone for transmitting images over wireless channels, where the VSSM-CA primarily consists of the generalized state space models (GSSM) and the zero-parameter, zero-computational channel adaptation method (CSI-ReST). We design the GSSM module, leveraging reversible matrix transformations to express generalized scan expanding operations, and theoretically prove that two GSSM modules can effectively capture global information. We discover that GSSM inherently possesses the ability to adapt to channels, a form of endogenous intelligence. Based on this, we design the CSI-ReST method, which injects channel state information (CSI) into the initial state of GSSM to utilize its native response, and into the residual state to mitigate CSI forgetting, enabling effective channel adaptation without introducing additional computational and parameter overhead. Experimental results on different devices, including IoT device JETSON AGX ORIN, show that MambaJSCC not only outperforms existing JSCC methods (e.g., SwinJSCC) across various scenarios but also significantly reduces parameter size, computational overhead, and inference delay. We have released our code and pre-trained models at https://github.com/Wireless3C-SJTU/MambaJSCC, allowing full reproduction of our results.