{"id":46291,"date":"2023-07-26T08:06:01","date_gmt":"2023-07-26T06:06:01","guid":{"rendered":"https:\/\/www.inovex.de\/?p=46291"},"modified":"2023-07-26T09:12:45","modified_gmt":"2023-07-26T07:12:45","slug":"addressing-vulnerabilities-in-federated-learning-through-blockchain","status":"publish","type":"post","link":"https:\/\/www.inovex.de\/de\/blog\/addressing-vulnerabilities-in-federated-learning-through-blockchain\/","title":{"rendered":"Addressing Vulnerabilities in Federated Learning through Blockchain"},"content":{"rendered":"

Federated learning allows a model to be learned from multiple participants without transferring the data to a central location. The approach promises to reduce privacy and latency risks but also has weaknesses such as data quality control and data security.\u00a0<\/span><\/p>\n

In this article, we propose integrating a voting protocol with a permissioned blockchain to address several federated learning vulnerabilities. We analyze the performance and robustness of the approach and discuss the results in a predictive maintenance use case. Basic knowledge of federated learning is assumed and can be, for example, looked up in <\/span>this blog post<\/span><\/a>.\u00a0<\/span><\/p>\n\n

\n

<\/p>\n<\/div>