20240054352. COMPUTER IMPLEMENTED MULTI-LAYER FEDERATED LEARNING METHOD BASED ON DISTRIBUTED CLUSTERING AND NON-TRANSITORY COMPUTER READABLE MEDIUM THEREOF simplified abstract (ASIA UNIVERSITY)
COMPUTER IMPLEMENTED MULTI-LAYER FEDERATED LEARNING METHOD BASED ON DISTRIBUTED CLUSTERING AND NON-TRANSITORY COMPUTER READABLE MEDIUM THEREOF
Organization Name
Inventor(s)
ZON-YIN Shae of TAIPEI CITY (TW)
KUN-YI Chen of TAICHUNG CITY (TW)
JING-PHA Tsai of TAICHUNG CITY (TW)
CHI-YU Chang of TAICHUNG CITY (TW)
YUAN-YU Tsai of TAICHUNG CITY (TW)
COMPUTER IMPLEMENTED MULTI-LAYER FEDERATED LEARNING METHOD BASED ON DISTRIBUTED CLUSTERING AND NON-TRANSITORY COMPUTER READABLE MEDIUM THEREOF - A simplified explanation of the abstract
This abstract first appeared for US patent application 20240054352 titled 'COMPUTER IMPLEMENTED MULTI-LAYER FEDERATED LEARNING METHOD BASED ON DISTRIBUTED CLUSTERING AND NON-TRANSITORY COMPUTER READABLE MEDIUM THEREOF
Simplified Explanation
The abstract describes a multi-layer federated learning method based on distributed clustering.
- Computing feature distribution similarity for participating nodes with non-iid data sets and grouping nodes into clusters based on this similarity.
- Updating local models of nodes in each cluster using a federated learning algorithm.
- Inputting nodes into a multi-layer aggregation mechanism.
- Terminating the aggregation mechanism when clustering result meets required demand.
- Implementing a blockchain-based multi-layer federated learning system with various modules.
- Potential Applications:**
- This technology can be applied in industries where data privacy is a concern, such as healthcare and finance. - It can also be used in IoT devices for collaborative learning without sharing raw data.
- Problems Solved:**
- Addresses the issue of data privacy by allowing nodes to learn collaboratively without sharing sensitive information. - Improves learning performance by utilizing distributed clustering to group nodes with similar data distributions.
- Benefits:**
- Enhanced data privacy protection. - Improved learning performance through distributed clustering. - Scalability and efficiency in federated learning processes.
Original Abstract Submitted
a multi-layer federated learning method based on distributed clustering is provided, which comprises the following steps. computing a feature distribution similarity for each of participating nodes with non-(non-independent and identically distributed) data sets, and grouping these nodes into plural clusters by the feature distribution similarity. updating local model of nodes of each cluster by a federated learning algorithm, and inputting these nodes into a multi-layer aggregation mechanism. terminating the operation of the multi-layer aggregation mechanism until the clustering result meets a required demand. furthermore, we implement a blockchain-based multi-layer federated learning system, including model aggregation module, api module, time-series synchronization module, and ipfs, based on distributed clustering architecture. the learning performance is proven to effectively improved.
