Federated Edge Intelligence for Next Generation AI Systems

Dear Colleagues,

In the context of 6G networks and advanced edge AI (artificial intelligence) applications, Federated Edge Intelligence (FEI) is rapidly becoming a key technology for realizing distributed, scalable, and privacy-preserving AI systems. FEI is a new type of distributed intelligence architecture that combines edge computing, federated learning (FL), and AI technologies and has remarkably advanced and huge application potential. Its main advantages lie in data privacy protection, low-latency response, and resource optimization. At the application level, FEI supports multi-modal data processing and can fuse visual, audio, tactile, radar, and other sensor data for real-time analysis and decision-making, representing significant application value in the fields of smart cities, autonomous driving, the industrial Internet, healthcare, financial economics, and other fields. Moreover, FEI’s edge computing architecture can be adaptively adjusted according to the computing power of different nodes, thus meeting real-time decision-making needs and providing more efficient computing and resource allocation. However, despite the many advantages of FEI, its realization still faces certain challenges. For example, edge nodes usually have more limited computing power and storage resources, and how to efficiently run complex AI models and conduct local training under these constraints remains an important issue. How to effectively synchronize models from different edge nodes and ensure the consistency of model aggregation is also an important challenge for FL in edge computing environments. 

We invite researchers to submit original, high-quality papers related to the above topics. All papers will undergo a rigorous peer review process. Submissions should clearly demonstrate innovative approaches, experimental results, and potential applications, particularly in advancing the field of Federated Edge Intelligence. 

This Special Issue covers the following broad topics related to FEI, specifically including, but not limited to, the following:

1. Federated learning at the edge:

• Model distribution and synchronization techniques in federated edge networks.
• Edge AI model optimization for real-time and end-device applications.
• Data upload/download mechanisms in federated learning.
• Model uploading and aggregation for local training. 

2. AI models in edge nodes:

• AI algorithms for multi-modal sensing in edge AI devices, including vision, audio, radar, touch, and smell.
• Edge node architectures and platform designs that support multiple AI models.
• Neuromorphic computing and pulsed neural networks (SNNs) in edge AI.

3. AI integration at the edge with the cloud:

• Task offloading techniques from edge devices to central cloud servers.
• Efficient communication and data-sharing protocols between edge nodes and cloud servers.
• Cloud model aggregation methods based on distributed edge node contributions.

4. Applications of Federated Edge Intelligence:

• Real-time applications such as autonomous driving, smart cities, the IoT, and healthcare.
• Support of intelligent edge computing for multi-sensor systems in 6G environments.
• Applications of Federated Edge Intelligence in industry.
• Federated learning in smart education.

Dr. Chunjiong Zhang
Dr. Weiwei Jiang
Dr. Tao Xie
Topic Editors