11-14, 13:10–13:20 (Asia/Bangkok), Stage 4
Federated learning enables multiple parties to contribute their locally trained models to an aggregation server, which securely combines individual models into a global one. However, it lacks a fair, verifiable, and proportionate reward (or penalty) mechanism for each contributor. Implementing a smart contract-based contribution analysis framework for federated learning on a privacy-enabled Ethereum L2 can address this challenge, and build the economics of federated learning public chain.
Sudhir heads the Engineering team Kinexys ( formerly Onyx), by J.P. Morgan. During his tenure at J.P. Morgan, Sudhir has been engaged in incubating, designing, developing and delivering several innovative products and solutions across multiple lines of business, including Risk Management, Trading, and Pricing, among others. Prior to joining J.P. Morgan, Sudhir was one of the early joiners at one of the most successful startups (BEA Systems) in Silicon Valley.