研究方向 | 边缘智能系统实验室

边缘智能系统实验室

云边协同的大模型服务

边缘侧算力有限，云端模型强但远、贵。我们研究如何在云和边之间分配请求、缓存和计算，让大模型服务的响应速度、吞吐和回答质量都更好。

Request SchedulingKV CacheSemantic CacheVector DatabaseReinforcement Learning

ICML 2026CCF-A

HeraSys: Collaborative Serving of Multiple LLM Workflows via Fine-Grained End-to-End Optimization

Size Li, Zhiqing Tang*, Hongrui Liang, Jianxiong Guo, Jiong Lou, Tian Wang, Weijia Jia

International Conference on Machine Learning (ICML), 2026

arXiv

ICML 2026CCF-A

Semantic Cache Distillation: Efficient State Transfer via Reuse and Selective Patching

Qianli Ma, Zhiqing Tang*, Hanshuai Cui, Zhi Yao, Weijia Jia*

International Conference on Machine Learning (ICML), 2026

arXiv

TSC 2025CCF-A广东省计算机学会优秀论文奖二等奖

Enhancing LLM QoS through Cloud-Edge Collaboration: A Diffusion-based Multi-Agent Reinforcement Learning Approach

Zhi Yao, Zhiqing Tang*, Wenmian Yang, Weijia Jia*

IEEE Transactions on Services Computing (TSC), 2025, 18(3): 1412–1427

DOI PDF

TSC 2026CCF-A

Adaptive Request Scheduling and Load Balancing for Edge Deployed Large Language Models

Fangyi Mou, Zhiqing Tang*, Weijia Jia*, Wei Zhao

IEEE Transactions on Services Computing (TSC), 2026

查看本方向全部论文 · 7 →

复杂大模型应用的端到端加速

真实的大模型应用是由 RAG、智能体和多个模型组成的工作流。我们研究如何在异构硬件上对整个工作流做端到端优化。

Agentic WorkflowsMulti-Model PipelinesBatchingPipelining

ICML 2026CCF-A

HeraSys: Collaborative Serving of Multiple LLM Workflows via Fine-Grained End-to-End Optimization

Size Li, Zhiqing Tang*, Hongrui Liang, Jianxiong Guo, Jiong Lou, Tian Wang, Weijia Jia

International Conference on Machine Learning (ICML), 2026

arXiv

TSC 2026CCF-A

Adapting Multi-Model Inference Pipelines with Diffusion-Based Reinforcement Learning in Edge Computing

Jinhao Sheng, Zhiqing Tang*, Jianxiong Guo, Kun Yue, Tian Wang, Weijia Jia

IEEE Transactions on Services Computing (TSC), 2026, Accepted

HPCC 2024CCF-C

Adaptive Configuration Selection for Multi-Model Inference Pipelines in Edge Computing

Jinhao Sheng, Zhiqing Tang*, Jianxiong Guo, Tian Wang

2024 The 26th IEEE International Conference on High Performance Computing and Communications (HPCC)

生成式 AI 的缓存与并行加速

图像和视频生成模型要迭代几十步去噪，非常慢。我们利用相邻步之间计算的冗余，通过特征缓存、特征预测和多设备协同，在几乎不损失质量的前提下大幅加速生成。

Diffusion TransformersFeature CachingVideo GenerationAIGC Scheduling

ICLR 2026CCF-A

BWCache: Accelerating Video Diffusion Transformers through Block-Wise Caching

Hanshuai Cui, Zhiqing Tang*, Zhifei Xu, Zhi Yao, Wenyi Zeng, Weijia Jia

International Conference on Learning Representations (ICLR), 2026

arXiv Code Project Page

ACM MM 2026CCF-A

Predict to Skip: Linear Multistep Feature Forecasting for Efficient Diffusion Transformers

Hanshuai Cui, Zhiqing Tang*, Qianli Ma, Zhi Yao, Weijia Jia*, Wei Zhao

ACM International Conference on Multimedia (ACM MM), 2026

arXiv Code Project Page

TMC 2026CCF-A

EAT: QoS-Aware Edge-Collaborative AIGC Task Scheduling via Attention-Guided Diffusion Reinforcement Learning

Zhifei Xu, Zhiqing Tang*, Jiong Lou, Zhi Yao, Xuan Xie, Tian Wang, Yinglong Wang, Weijia Jia

IEEE Transactions on Mobile Computing (TMC), 2026, Accepted

arXiv Code

ACM MM 2026CCF-A

Not All Frames Deserve Full Computation: Accelerating Autoregressive Video Generation via Selective Computation and Predictive Extrapolation

Hanshuai Cui, Zhiqing Tang*, Zhi Yao, Fanshuai Meng, Weijia Jia, Wei Zhao

ACM International Conference on Multimedia (ACM MM), 2026

arXiv Code Project Page

边缘容器调度与资源管理

边缘集群资源紧张、负载波动大。我们研究感知镜像层的容器调度、迁移与更新，冷启动优化和资源超售，并用深度强化学习和在线算法做决策。

Layer-Aware SchedulingKubernetesServerlessDeep Reinforcement LearningOnline Algorithms

TMC 2026CCF-A

EAT: QoS-Aware Edge-Collaborative AIGC Task Scheduling via Attention-Guided Diffusion Reinforcement Learning

Zhifei Xu, Zhiqing Tang*, Jiong Lou, Zhi Yao, Xuan Xie, Tian Wang, Yinglong Wang, Weijia Jia

IEEE Transactions on Mobile Computing (TMC), 2026, Accepted

arXiv Code

MSN 2024CCF-C最佳论文奖

LRScheduler: A Layer-aware and Resource-adaptive Container Scheduler in Edge Computing

Zhiqing Tang, Wentao Peng, Jianxiong Guo, Jiong Lou, Hanshuai Cui, Tian Wang, Yuan Wu, Weijia Jia

2024 20th International Conference on Mobility, Sensing and Networking (MSN)

Code arXiv

TMC 2023CCF-A

Layer Dependency-aware Learning Scheduling Algorithms for Containers in Mobile Edge Computing

Zhiqing Tang, Jiong Lou, Weijia Jia

IEEE Transactions on Mobile Computing (TMC), 2023, 22(6): 3444–3459

DOI PDF

TMC 2026CCF-A

Hybrid Learning for Cold-Start-Aware Microservice Scheduling in Dynamic Edge Environments

Jingxi Lu, Wenhao Li, Jianxiong Guo, Xingjian Ding, Zhiqing Tang, Tian Wang, Weijia Jia

IEEE Transactions on Mobile Computing (TMC), 2026, Accepted

查看本方向全部论文 · 25 →

高效多模态推理与检索

图像、文本、音频、视频混合的检索和问答，既要语义对得准，又要在边缘设备上算得快、存得下。

Multimodal EmbeddingVector RetrievalMultimodal RAGFederated Learning

TOIT 2025CCF-B

Optimizing Communication Efficiency through Training Potential in Multi-Modal Federated Learning

Yinghao Zhang, Jianxiong Guo, Xingjian Ding, Zhiqing Tang, Tian Wang, Weili Wu, Weijia Jia

ACM Transactions on Internet Technology, 2025, 25(3): 1–31

边缘计算中的任务调度与优化

课题组长期研究边缘计算中的任务卸载与调度、服务部署、无人机与车联网边缘计算、众包与联邦学习等问题，并撰写了端侧 AI 模型的综述。

Task OffloadingUAVVehicular NetworksFederated Learning

CSUR 2025JCR-Q1

Empowering Edge Intelligence: A Comprehensive Survey on On-Device AI Models

Xubin Wang, Zhiqing Tang, Jianxiong Guo, Tianhui Meng, Chenhao Wang, Tian Wang, Weijia Jia

ACM Computing Surveys (CSUR), 2025, 57(9): 1–39

DOI PDF

TOIT 2025CCF-B

Optimizing Communication Efficiency through Training Potential in Multi-Modal Federated Learning

Yinghao Zhang, Jianxiong Guo, Xingjian Ding, Zhiqing Tang, Tian Wang, Weili Wu, Weijia Jia

ACM Transactions on Internet Technology, 2025, 25(3): 1–31

TBD 2025JCR-Q1

Look Closer to Your Enemy: Learning to Attack via Teacher-Student Mimicking

Mingjie Wang, Jianxiong Guo*, Sirui Li, Dingwen Xiao, Zhiqing Tang*

IEEE Transactions on Big Data (TBD), 2025, Early Access

DOI PDF

TMC 2025CCF-A

Blockchain-enabled Multiple Sensitive Task-offloading Mechanism for MEC Applications

Yang Xu, Hangfan Li, Cheng Zhang, Zhiqing Tang, Xiaoxiong Zhong, Ju Ren, Hongbo Jiang, Yaoxue Zhang

IEEE Transactions on Mobile Computing (TMC), 2025, 24(4): 3241–3255

DOI PDF

查看本方向全部论文 · 19 →