Jian Wang
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
Jian Wang has a total of 24 co-authors. Here are the top co-authors by number of shared publications:
- Gang Feng
National Key Laboratory of Wireless Communications, University of Electronic Science and Technology of China, Chengdu, China
5 shared publications - Shuang Qin
National Key Laboratory of Science and Technology on Communications, University of Electronic Science and Technology of China, Chengdu, China
4 shared publications - Lei Cheng
National Key Laboratory of Wireless Communications, University of Electronic Science and Technology of China, Chengdu, China
3 shared publications - Yijing Liu
National Key Laboratory of Wireless Communications, University of Electronic Science and Technology of China, Chengdu, China
3 shared publications - Xinyi Xu
National Key Laboratory of Wireless Communications, University of Electronic Science and Technology of China, Chengdu, China
2 shared publications - Yanan Liu
School of Network Security, Jinling Institute of Technology, Nanjing 211169, China
2 shared publications
Latest Publications
- 10 September 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Mobile Computing
- Vol. 25 (2) , 2691-2706
Joint Inference Offloading and Model Caching for Small and Large Language Model Collaboration
Large Language Models (LLMs), with advanced content creation and inference capabilities, can provide immersive intelligent services to users in mobile edge networks. However, the increasing demand for real-time artificial intelligence (AI) applications aggravates the limitations of cloud-based LLMs due to the long response time. Meanwhile, Small Language Models (SLMs), which are cost-effective and locally deployable for terminal devices, can serve as an efficient supplement to LLMs for performing latency-sensitive tasks with lower generalization capability. Due to the resource constraints of edge networks and the diverse requirements of user tasks, it is critical to design an inference framework that effectively coordinates the deployment and collaboration of LLMs and SLMs. In this paper, we propose an LLM-SLM collaborative inference (LSCI) scheme under a mobile edge computing (MEC) architecture, which jointly decides where to cache models and how to offload inference tasks to balance latency, accuracy, and resource costs. To optimize inference performance subject to resource constraints, we jointly solve the inference task offloading and model caching problem in LSCI scheme. Specifically, we employ deep reinforcement learning (DRL) to select highly popular SLMs to be cached on the edge server, and distributed belief propagation technique to solve the associated inference task offloading issue. Numerical results show that the proposed LSCI scheme can achieve significant performance gain in terms of inference performance when compared with a number of baseline solutions.
- 10 September 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Mobile Computing
- Vol. 25 (2) , 2691-2706
Joint Inference Offloading and Model Caching for Small and Large Language Model Collaboration
Large Language Models (LLMs), with advanced content creation and inference capabilities, can provide immersive intelligent services to users in mobile edge networks. However, the increasing demand for real-time artificial intelligence (AI) applications aggravates the limitations of cloud-based LLMs due to the long response time. Meanwhile, Small Language Models (SLMs), which are cost-effective and locally deployable for terminal devices, can serve as an efficient supplement to LLMs for performing latency-sensitive tasks with lower generalization capability. Due to the resource constraints of edge networks and the diverse requirements of user tasks, it is critical to design an inference framework that effectively coordinates the deployment and collaboration of LLMs and SLMs. In this paper, we propose an LLM-SLM collaborative inference (LSCI) scheme under a mobile edge computing (MEC) architecture, which jointly decides where to cache models and how to offload inference tasks to balance latency, accuracy, and resource costs. To optimize inference performance subject to resource constraints, we jointly solve the inference task offloading and model caching problem in LSCI scheme. Specifically, we employ deep reinforcement learning (DRL) to select highly popular SLMs to be cached on the edge server, and distributed belief propagation technique to solve the associated inference task offloading issue. Numerical results show that the proposed LSCI scheme can achieve significant performance gain in terms of inference performance when compared with a number of baseline solutions.
- 8 August 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Network Science and Engineering
- Vol. 13, 1569-1585
Split Learning Based Cloud-Edge-End Collaborative Model Training in Heterogeneous Networks
Large language models (LLMs) demonstrate significant potential for enabling intelligent endogenous networks owing to their amazing intelligence level. LLMs are currently deployed on cloud servers as their vast parameter scales introduce a substantial computational burden to pre-training and inference processes. This deployment paradigm faces severe challenges, including insufficient personalization, high inference latency, and privacy concerns, as an increasing number of mobile users enjoy the LLM services. To enhance the personalization of pre-trained LLMs, exploiting massive local datasets distributed across edge devices to fine-tune LLMs is essential. In this paper, we propose a split learning-based cloud-edge-end collaborative training framework (SCCT) to harness the abundant computational resources of cloud and edge servers while exploiting massive distributed datasets on edge devices for LLM fine-tuning. In SCCT, an LLM with a parameter-efficient fine-tuning module is deployed on the cloud server, while a small-scale language model (SLM) is deployed across each edge device participating in SCCT and its associated edge server in the manner of split learning. SLM and LLM collaborate in a serial manner for training and inference. To optimize the collaboration efficiency, we formulate a mixed integer nonlinear programming problem to minimize the training latency of SCCT, considering the resource heterogeneity of edge devices and network dynamics. To solve this problem, a two-timescale optimization algorithm is proposed to determine the optimal split points on the large timescale and the allocation of communication and computational resources of the edge server on the small timescale. Numerical results demonstrate that the proposed two-timescale optimization algorithm assisted SCCT outperforms the state-of-the-art LLMs deployment and training paradigm in terms of training efficiency and inference performance.
- 8 August 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Network Science and Engineering
- Vol. 13, 1569-1585
Split Learning Based Cloud-Edge-End Collaborative Model Training in Heterogeneous Networks
Large language models (LLMs) demonstrate significant potential for enabling intelligent endogenous networks owing to their amazing intelligence level. LLMs are currently deployed on cloud servers as their vast parameter scales introduce a substantial computational burden to pre-training and inference processes. This deployment paradigm faces severe challenges, including insufficient personalization, high inference latency, and privacy concerns, as an increasing number of mobile users enjoy the LLM services. To enhance the personalization of pre-trained LLMs, exploiting massive local datasets distributed across edge devices to fine-tune LLMs is essential. In this paper, we propose a split learning-based cloud-edge-end collaborative training framework (SCCT) to harness the abundant computational resources of cloud and edge servers while exploiting massive distributed datasets on edge devices for LLM fine-tuning. In SCCT, an LLM with a parameter-efficient fine-tuning module is deployed on the cloud server, while a small-scale language model (SLM) is deployed across each edge device participating in SCCT and its associated edge server in the manner of split learning. SLM and LLM collaborate in a serial manner for training and inference. To optimize the collaboration efficiency, we formulate a mixed integer nonlinear programming problem to minimize the training latency of SCCT, considering the resource heterogeneity of edge devices and network dynamics. To solve this problem, a two-timescale optimization algorithm is proposed to determine the optimal split points on the large timescale and the allocation of communication and computational resources of the edge server on the small timescale. Numerical results demonstrate that the proposed two-timescale optimization algorithm assisted SCCT outperforms the state-of-the-art LLMs deployment and training paradigm in terms of training efficiency and inference performance.
- 24 June 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Vehicular Technology
- Vol. 74 (11) , 18096-18111
Access Control in Satellite–Terrestrial Backhaul Network: An Emergent Communication Based Multi-Agent Reinforcement Learning Approach
The satellite-terrestrial backhaul network (STBN) has recently emerged as a promising architecture to provide cost-effective seamless communication services to mobile users, especially in remote areas. In such networks, limited backhaul capacity, dynamic channel states, and user mobility pose significant challenges to user access control, involving efficient channel allocation and interference coordination to maximize user quality of service and network resource utilization. This dynamic user access problem can be formulated as a sequential decision process, and thus multi-agent reinforcement learning (MARL) can be a potential solution. However, conventional MARL lacks cooperation mechanisms for accurate state inference, which is critical for making optimal and cooperative resource allocation and interference management actions in such a dynamic and resource-constrained STBN. Moreover, although centralized coordination may improve system performance, it raises privacy and scalability concerns. In this paper, we propose a decentralized cooperative user access strategy, by leveraging emergent communication based MARL (EC-MARL). EC-MARL incorporates an emergent communication mechanism into the MARL framework, enabling user equipments (UEs) to cooperatively infer the network state more accurately while preserving privacy and scalability. Specifically, each agent encodes its local observation into a cooperation message, and shares the message among other agents. Based on these shared messages and their own local observations, UEs cooperatively make access decisions. Simulation results demonstrate that the proposed EC-MARL approach significantly outperforms several benchmarks in various network settings.
- 24 June 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Vehicular Technology
- Vol. 74 (11) , 18096-18111
Access Control in Satellite–Terrestrial Backhaul Network: An Emergent Communication Based Multi-Agent Reinforcement Learning Approach
The satellite-terrestrial backhaul network (STBN) has recently emerged as a promising architecture to provide cost-effective seamless communication services to mobile users, especially in remote areas. In such networks, limited backhaul capacity, dynamic channel states, and user mobility pose significant challenges to user access control, involving efficient channel allocation and interference coordination to maximize user quality of service and network resource utilization. This dynamic user access problem can be formulated as a sequential decision process, and thus multi-agent reinforcement learning (MARL) can be a potential solution. However, conventional MARL lacks cooperation mechanisms for accurate state inference, which is critical for making optimal and cooperative resource allocation and interference management actions in such a dynamic and resource-constrained STBN. Moreover, although centralized coordination may improve system performance, it raises privacy and scalability concerns. In this paper, we propose a decentralized cooperative user access strategy, by leveraging emergent communication based MARL (EC-MARL). EC-MARL incorporates an emergent communication mechanism into the MARL framework, enabling user equipments (UEs) to cooperatively infer the network state more accurately while preserving privacy and scalability. Specifically, each agent encodes its local observation into a cooperation message, and shares the message among other agents. Based on these shared messages and their own local observations, UEs cooperatively make access decisions. Simulation results demonstrate that the proposed EC-MARL approach significantly outperforms several benchmarks in various network settings.
- 13 June 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Cognitive Communications and Networking
- Vol. 12, 1211-1226
Scheduling Model Transmission and Training for Contact-Based Decentralized Satellite Federated Learning
The significant enhancement of satellite onboard processing capability and drastic proliferation of Artificial Intelligence (AI) applications have fostered decentralized satellite federated learning (DSFL), a transformative paradigm that exchanges and aggregates machine learning (ML) models in satellite clusters for collaborative learning. However, the limited model exchange opportunities caused by intermittent inter-satellite contacts, along with heterogeneous onboard datasets, can lead to ineffective and/or biased model aggregation. To address these issues, it is crucial yet challenging to design an effective DSFL scheduling strategy that determines whether and when to pull models from contact satellites and perform local training for optimizing DSFL performance. In this paper, we propose a contact-based DSFL framework and formulate the DSFL scheduling problem to maximize the accuracy of trained models. As the problem cannot be solved directly, we transform it into a hierarchical Markov game by introducing options for decision agents deployed on individual satellites. Under a learning-to-learn paradigm, we develop a Multi-agent Dueling Double Deep Q Network (MA3DQN)-based intelligent DSFL scheduling strategy. The agents, trained in a distributed and alternating manner, adaptively make scheduling decisions based on instantaneous partial observations of the environment. Simulation results demonstrate the efficiency and adaptability of the MA3DQN-based strategy over three baselines.
- 13 June 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Cognitive Communications and Networking
- Vol. 12, 1211-1226
Scheduling Model Transmission and Training for Contact-Based Decentralized Satellite Federated Learning
The significant enhancement of satellite onboard processing capability and drastic proliferation of Artificial Intelligence (AI) applications have fostered decentralized satellite federated learning (DSFL), a transformative paradigm that exchanges and aggregates machine learning (ML) models in satellite clusters for collaborative learning. However, the limited model exchange opportunities caused by intermittent inter-satellite contacts, along with heterogeneous onboard datasets, can lead to ineffective and/or biased model aggregation. To address these issues, it is crucial yet challenging to design an effective DSFL scheduling strategy that determines whether and when to pull models from contact satellites and perform local training for optimizing DSFL performance. In this paper, we propose a contact-based DSFL framework and formulate the DSFL scheduling problem to maximize the accuracy of trained models. As the problem cannot be solved directly, we transform it into a hierarchical Markov game by introducing options for decision agents deployed on individual satellites. Under a learning-to-learn paradigm, we develop a Multi-agent Dueling Double Deep Q Network (MA3DQN)-based intelligent DSFL scheduling strategy. The agents, trained in a distributed and alternating manner, adaptively make scheduling decisions based on instantaneous partial observations of the environment. Simulation results demonstrate the efficiency and adaptability of the MA3DQN-based strategy over three baselines.
- 2 August 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Vehicular Technology
- Vol. 73 (12) , 18846-18862
Intelligent Beam Configuration for Neighbor Discovery in Ad Hoc Networks With Directional Antennas
High frequency directional communication is considered as key technology to improve the performance of mobile Ad Hoc networks owing to its advantages in terms of communication distance and low interference. Neighbor discovery plays a key role for efficient routing and topology control in mobile Ad Hoc networks. It is also a very challenging issue due to the use of directional antennas and the movement of mobile nodes. Beam configuration is the core step of neighbor discovery in mobile Ad Hoc networks with directional antennas. Therefore, it is imperative to develop an efficient beam configuration algorithm to reduce the neighbor discovery latency. In this paper, we propose a novel beam configuration algorithm based on personalized federated learning. Considering the characteristics of mobile Ad Hoc networks (e.g., dynamic topology and directional communication), we use Deep Deterministic Policy Gradient (DDPG) as the local model of federated learning. Since the local data of Ad Hoc network nodes is heterogeneous, Model Agnostic Meta Learning (MAML) is applied to personalize the federated learning. Numerical results demonstrate that our proposed algorithm has better performance than some baseline algorithms
- 2 August 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Vehicular Technology
- Vol. 73 (12) , 18846-18862
Intelligent Beam Configuration for Neighbor Discovery in Ad Hoc Networks With Directional Antennas
High frequency directional communication is considered as key technology to improve the performance of mobile Ad Hoc networks owing to its advantages in terms of communication distance and low interference. Neighbor discovery plays a key role for efficient routing and topology control in mobile Ad Hoc networks. It is also a very challenging issue due to the use of directional antennas and the movement of mobile nodes. Beam configuration is the core step of neighbor discovery in mobile Ad Hoc networks with directional antennas. Therefore, it is imperative to develop an efficient beam configuration algorithm to reduce the neighbor discovery latency. In this paper, we propose a novel beam configuration algorithm based on personalized federated learning. Considering the characteristics of mobile Ad Hoc networks (e.g., dynamic topology and directional communication), we use Deep Deterministic Policy Gradient (DDPG) as the local model of federated learning. Since the local data of Ad Hoc network nodes is heterogeneous, Model Agnostic Meta Learning (MAML) is applied to personalize the federated learning. Numerical results demonstrate that our proposed algorithm has better performance than some baseline algorithms
- Open Access
- 2 April 2015
- Published by Springer Nature in BMC Cancer
- Vol. 15 (1) , 213
Sentinel lymph node detection using magnetic resonance lymphography with conventional gadolinium contrast agent in breast cancer: a preliminary clinical study
Sentinel lymph node (SLN) mapping is the standard method for axillary lymph node staging in patients with breast cancer. Blue dye and radioisotopes are commonly used agents to localize SLNs, but both have several disadvantages. The purpose of this study was to evaluate magnetic resonance lymphography with a gadolinium-based contrast agent (Gd-MRL) in sentinel lymph node identification and metastasis detection in patients with breast cancer. Sixty patients (mean age: 46.2 ± 8.8 years) with stage T1- 2 breast cancer and clinically negative axillary lymph nodes participated in this study. After 0.9 ml of contrast material and 0.1 ml of mepivacaine hydrochloride 1% were mixed and injected intradermally into the upper-outer periareolar areas, axillary lymph flow was tracked and sentinel lymph nodes were identified by Gd-MRL. After SLN biopsy and/or surgery, the efficacy of SLN identification and metastasis detection of Gd-MRL were examined. Ninety-six lymph nodes were identified by Gd-MRL as SLNs (M-SLN), and 135 lymph nodes were detected by blue dye-guided methods as SLNs (D-SLN). There was a strong correlation (P < 0.001) between the SLN numbers found by these two methods. Using blue dye-guided methods as the gold standard, the sensitivity of Gd-MRL was 95.65% and the false-negative rate was 4.3% for axillary lymphatic metastasis detection. With heterogeneous enhancement and enhancement defect as the diagnostic criteria, Gd-MRL gave a sensitivity of 89.29% and specificity of 89.66% in discriminating malignant from benign SLNs. Gd-MRL offers a new method for SLN identification and metastasis detection in patients with breast cancer.
- 2 April 2015
- Published by Springer Nature in BMC Cancer
- Vol. 15 (1) , 213
Open AccessSentinel lymph node detection using magnetic resonance lymphography with conventional gadolinium contrast agent in breast cancer: a preliminary clinical study
Sentinel lymph node (SLN) mapping is the standard method for axillary lymph node staging in patients with breast cancer. Blue dye and radioisotopes are commonly used agents to localize SLNs, but both have several disadvantages. The purpose of this study was to evaluate magnetic resonance lymphography with a gadolinium-based contrast agent (Gd-MRL) in sentinel lymph node identification and metastasis detection in patients with breast cancer. Sixty patients (mean age: 46.2 ± 8.8 years) with stage T1- 2 breast cancer and clinically negative axillary lymph nodes participated in this study. After 0.9 ml of contrast material and 0.1 ml of mepivacaine hydrochloride 1% were mixed and injected intradermally into the upper-outer periareolar areas, axillary lymph flow was tracked and sentinel lymph nodes were identified by Gd-MRL. After SLN biopsy and/or surgery, the efficacy of SLN identification and metastasis detection of Gd-MRL were examined. Ninety-six lymph nodes were identified by Gd-MRL as SLNs (M-SLN), and 135 lymph nodes were detected by blue dye-guided methods as SLNs (D-SLN). There was a strong correlation (P < 0.001) between the SLN numbers found by these two methods. Using blue dye-guided methods as the gold standard, the sensitivity of Gd-MRL was 95.65% and the false-negative rate was 4.3% for axillary lymphatic metastasis detection. With heterogeneous enhancement and enhancement defect as the diagnostic criteria, Gd-MRL gave a sensitivity of 89.29% and specificity of 89.66% in discriminating malignant from benign SLNs. Gd-MRL offers a new method for SLN identification and metastasis detection in patients with breast cancer.
- Open Access
- 1 September 2013
- Published by Institution of Engineering and Technology (IET) in IET Wireless Sensor Systems
- Vol. 3 (3) , 172-182
Intra‐cluster key sharing in hierarchical sensor networks
In hierarchical sensor networks, key sharing at the intra‐cluster tier (i.e. how to establish pairwise keys between a sensor and its cluster head) is more challenging than that at inter‐cluster tier. Despite some key management protocols have been proposed, there still lacks both secure and efficient one. In this study, the authors proposed an intra‐cluster key sharing to overcome the security problems nowadays. The analysis proved that this proposed scheme has the following advantages by comparing with other schemes: (i) it is more efficient in saving the storage overheads and communication overheads; (ii) it improves the resilience against node capture attacks during and after initialisation; and (iii) it guarantees 100% connectivity at intra‐cluster communication tier.
- 1 September 2013
- Published by Institution of Engineering and Technology (IET) in IET Wireless Sensor Systems
- Vol. 3 (3) , 172-182
Open AccessIntra‐cluster key sharing in hierarchical sensor networks
In hierarchical sensor networks, key sharing at the intra‐cluster tier (i.e. how to establish pairwise keys between a sensor and its cluster head) is more challenging than that at inter‐cluster tier. Despite some key management protocols have been proposed, there still lacks both secure and efficient one. In this study, the authors proposed an intra‐cluster key sharing to overcome the security problems nowadays. The analysis proved that this proposed scheme has the following advantages by comparing with other schemes: (i) it is more efficient in saving the storage overheads and communication overheads; (ii) it improves the resilience against node capture attacks during and after initialisation; and (iii) it guarantees 100% connectivity at intra‐cluster communication tier.
- Open Access
- 20 August 2011
- Published by China Science Publishing & Media Ltd. in Journal of Electronics & Information Technology
- Vol. 33 (8) , 1913-1918
Threshold Key Sharing Model in Wireless Sensor Networks
- 20 August 2011
- Published by China Science Publishing & Media Ltd. in Journal of Electronics & Information Technology
- Vol. 33 (8) , 1913-1918
Open AccessThreshold Key Sharing Model in Wireless Sensor Networks