Research Article
Hashgraph Based Federated Learning for Secure Data Sharing
@INPROCEEDINGS{10.1007/978-3-030-69072-4_45, author={Xiuxian Zhang and Lingyu Zhao and Jinfeng Li and Xiaorong Zhu}, title={Hashgraph Based Federated Learning for Secure Data Sharing}, proceedings={Wireless and Satellite Systems. 11th EAI International Conference, WiSATS 2020, Nanjing, China, September 17-18, 2020, Proceedings, Part II}, proceedings_a={WISATS PART 2}, year={2021}, month={2}, keywords={Hashgraph Federated learning Blockchain Gossip Virtual voting}, doi={10.1007/978-3-030-69072-4_45} }- Xiuxian Zhang
Lingyu Zhao
Jinfeng Li
Xiaorong Zhu
Year: 2021
Hashgraph Based Federated Learning for Secure Data Sharing
WISATS PART 2
Springer
DOI: 10.1007/978-3-030-69072-4_45
Abstract
As the key technology of connected intelligence, the importance of artificial intelligence has increased rapidly. It is worth to note that the most critical challenge is the secure data sharing which is stored in different area and belonged to different organization. With this in mind, a hashgraph based federated learning for secure data sharing model is proposed to protect user privacy and detect the dishonest model provider. In terms of technologies, detection of the local model is added to the hashgraph consensus processing, and only if the supermajority (more than 2/3) of the participants agree, the local model could be adopted. Therefore, the accuracy and convergence rate of the federated learning both increased largely. On the other hand, the asynchronous working mode of hashgraph can greatly reduce network overload. Simulation results show that the hashgraph based federated learning enables the data sharing more secure and reliable.
