perspectiva aiqbenchSiguranță AI

Google is shifting the computational burden of federated learning from client devices to attested server-side Trusted Execution Environments (TEEs) for its Gboard keyboard. By publishing access policies to Sigstore's Rekor log and ensuring reproducible builds, this approach enables external verification of central differential privacy guarantees. This architectural change strengthens the trust model for on-device AI training by making privacy mechanisms auditable rather than opaque.

Surse care acoperăMarkTechPost, Latent Space

Ce este aceastaLectura unui model asupra articolului și a acoperirii conexe, scrisă pentru aiqbench. Nu sunt cuvintele publicației.

Google Research Moves Federated Learning Into TEEs: Gboard Now Trains With Externally Verifiable Differential Privacy

MarkTechPost4 octombrie 20261 min de cititArticolul și perspectiva sunt în engleză.

Google Research has unveiled a federated learning system that moves gradient computation from phones into attested server-side TEEs. Access policies are published to Sigstore's Rekor log and the binaries are reproducibly buildable, so central differential privacy can be checked externally. Gboard already uses it for English and Japanese next-word prediction. The post Google Research Moves Federated Learning Into TEEs: Gboard Now Trains With Externally Verifiable Differential Privacy appeared first on MarkTechPost.

Citește articolul complet pe MarkTechPost

Citește pe MarkTechPost ↗Întreabă despre acest articol