AWS – AWS Clean Rooms ML supports privacy-enhanced model training and inferencing

Today, AWS announces AWS Clean Rooms ML custom modeling, which enables organizations to generate predictive insights with their partners running their own machine-learning (ML) models and using their data in a clean rooms collaboration. With this launch, companies and their partners can train ML models and run inference on collective datasets without having to share sensitive data or proprietary models.

For example, advertisers can bring their proprietary model and data into a Clean Rooms collaboration, and invite publishers to join their data to train and deploy a custom ML model that helps them increase campaign effectiveness—all without sharing their custom model and data with one another. Similarly, financial institutions can use historical transaction records to train a custom ML model, and invite partners into a Clean Rooms collaboration to detect potential fraudulent transactions, without having to share underlying data and model among collaborators. With AWS Clean Rooms ML custom modeling, you can gain valuable insights with your partners while applying privacy-enhancing controls when running model training and inferencing by specifying the datasets to be used in a Clean Rooms environment. This allows you and your partners to approve the datasets used, and removes the need to share sensitive data or proprietary models with one another. AWS Clean Rooms ML also offers an AWS-authored lookalike modeling capability that can help you improve lookalike segment accuracy by up to 36% compared to industry baselines.

AWS Clean Rooms ML is available as a capability of AWS Clean Rooms in these AWS Regions. To learn more, visit AWS Clean Rooms ML.

Read More for the details.