privacy

Collection of privacy specs and discussion

• MURRE: Mechanism for User Reports with Regulated Epsilon. This proposal is a means to provide differentially private vectors for machine learning.
• SPURFOWL: Sandboxed Private User Reporting Functions Operating Within Limits. This proposal suggests that browsers maintain a trail store, run private sandboxed functions on them and then report them.
• COWBIRD: Coordinated Optimization Without Big Resource Demands. This document proposes browsers support a customizable federated learning platform for tiny machine learning models.
• PAURAQUE: Profile-based API for User Attribute Queries. Aims to solve for attribute-based targeting in a privacy preserving way.