A5060868822- Indoor and Outdoor Localization Technologies
- Privacy-Preserving Technologies in Data
- Target Tracking and Data Fusion in Sensor Networks
- Single-cell and spatial transcriptomics
- Optical Systems and Laser Technology
- Infrared Target Detection Methodologies
- Collaboration in agile enterprises
- Ultrasonics and Acoustic Wave Propagation
- Robotics and Sensor-Based Localization
- Winter Sports Injuries and Performance
- Advanced Graph Neural Networks
- Cancer-related molecular mechanisms research
- Underwater Vehicles and Communication Systems
- Mobile Crowdsensing and Crowdsourcing
- Distributed Sensor Networks and Detection Algorithms
- Diverse Approaches in Healthcare and Education Studies
- Domain Adaptation and Few-Shot Learning
- Stochastic Gradient Optimization Techniques
- Innovation in Digital Healthcare Systems
- Non-Destructive Testing Techniques
- Welding Techniques and Residual Stresses
- Cell Image Analysis Techniques
University of Science and Technology Beijing
2022-2025
Beijing Institute of Technology
2025
Tsinghua University
2024
Hangzhou Dianzi University
2023-2024
Tokyo Institute of Technology
2024
Ministry of Education of the People's Republic of China
2023
As a promising learning paradigm integrating computation and communication, federated (FL) proceeds the local training periodic sharing from distributed clients. Due to non-i.i.d. data distribution on clients, FL model suffers gradient diversity, poor performance, bad convergence, etc. In this work, we aim tackle key issue by adopting importance sampling (IS) for training. We propose (ISFL), an explicit framework with theoretical guarantees. Firstly, derive convergence theorem of ISFL...
Federated Learning (FL) is a promising distributed learning mechanism that revolutionizes our interaction with data in the IoT ecosystem. Due to rapidly growing scale of smart devices and limited transmission resources networks, simple, consistent scalable FL framework aiming address communication bottleneck urgently needed. In this work, we propose an efficient approach Selective Aggregation Models (SAM) mitigate overload systems. The introduction SAM enables each local client upload its...
In this paper, we introduce a novel cooperative target tracking algorithm, namely the quantum-inspired belief propagation, aimed at rectifying limitations observed in existing localization algorithms employed multi-target scenarios. Leveraging principles of quantum superposition, our algorithm seeks to alleviate uncertainty inherent message fusion within propagation frameworks, thereby enhancing accuracy and stability localization. The utilization Monte Carlo method facilitates simulation...
Abstract Trajectory inference is a crucial task in single-cell RNA-sequencing downstream analysis, which can reveal the dynamic processes of biological development, including cell differentiation. Dimensionality reduction an important step trajectory process. However, most existing methods rely on features derived from traditional dimensionality methods, such as principal component analysis and uniform manifold approximation projection. These are not specifically designed for fail to fully...
Cooperative localization is essential for many Internet of Things (IoT)-related applications in harsh environments. Generally, the inertial navigation system self-contained and adopted as basis a cooperative tracking system, but it still faces problem accumulated errors cannot provide long-term, high-precision positioning. The particle filter (PF) widely used to fuse multiple information inhibit accumulative errors. However, degradation impoverishment remain unsolved. This article proposed...
Real-time high-accuracy localization has a wide range of applications in scenarios, such as pedestrian navigation, emergency rescue, and vehicle networks. In these conditions, the measurement models are often nonlinear, traditional Kalman particle filters cannot provide long-time high-precision location-based services. To this end, we propose Gaussian condensation filter (GCF) algorithm that can achieve harsh environment. However, aiming at degradation sampling points target tracking based...
Recent advancements in reinforcement learning have made significant impacts across various domains, yet they often struggle complex multi-agent environments due to issues like algorithm instability, low sampling efficiency, and the challenges of exploration dimensionality explosion. Hierarchical (HRL) offers a structured approach decompose tasks into simpler sub-tasks, which is promising for settings. This paper advances field by introducing hierarchical architecture that autonomously...