Synaptics CPO Advocates for Federated Machine Learning at Edge

Vikram Gupta, Chief Product Officer at Synaptics, is advocating for the adoption of federated machine learning (FML) to transition artificial intelligence from centralized cloud processing to the edge. Gupta argues that traditional centralized training is becoming impractical because of latency, regulatory pressures, and privacy concerns.

Under the FML model, devices such as wearables, smart home assistants, and industrial sensors can become more autonomous. These devices train on local data and share only encrypted model updates rather than raw data, reducing the need for centralized data transmission.

Gupta acknowledges that the current edge environment is fragmented by diverse operating systems and chip architectures. He asserts that the future of edge AI depends on the use of modular hardware, open-source tools, and a multi-tiered intelligence approach that spans on-device processing, near-edge hubs, and the cloud.