Jing-pha Tsai

Title: Jing-pha Tsai: Inventor in Federated Learning

Introduction

Jing-pha Tsai is an inventor based in Taichung, Taiwan. His work includes a patented approach to multi-layer federated learning that uses distributed clustering to organize participating nodes.

Latest Patents

Tsai is associated with 1 patent: “Computer Implemented Multi-Layer Federated Learning Method Based on Distributed Clustering and Non-Transitory Computer Readable Medium Thereof.” The method calculates feature-distribution similarity among participating nodes with non-independent and identically distributed data, then groups the nodes into clusters. It updates each cluster’s local models through federated learning and uses a multi-layer aggregation mechanism, continuing until the clustering results meet specified requirements. The described system also incorporates blockchain-based components, including model aggregation, an API, time-series synchronization, and IPFS. The patent states that this architecture can improve learning performance.

Career Highlights

Tsai is associated with Asia University. His patent focuses on coordinating distributed data and model aggregation in federated learning.

Collaborations

His listed coworkers include Zon-yin Shae and Kun-yi Chen.

Conclusion

Jing-pha Tsai’s patented work applies distributed clustering and multi-layer aggregation to federated learning, connecting his research with Asia University and his listed collaborators.