AWS Certification Exam AI Practitioner (Practice Questions) set16

問題 14 / 65

I want to improve a model without sharing highly sensitive transaction data across multiple planets. Which distributed learning strategy is most appropriate?

Collect data from all planets and perform centralized training at headquarters.
Encrypt the raw data, transfer it, decrypt it centrally, and train.
Use federated learning that performs local training on each planet and updates the global model via secure aggregation.
Publish and share anonymized raw data.
Perform all updates using only synthetic data.

当サイトでは、ユーザー体験の向上を目的としてCookieを使用しています。サイトの利用を継続することで、Cookieの使用に同意したものとみなされます。