Skip to content

Anonymizing client data via simulated clients #2

Description

@KyleBenson

Increase the k in k-anonymity for the volunteers' privacy by modeling sensor clients from real data published by real clients and creating new "simulated clients" that randomly generate data from distributions learned by these models. This may take a week or more to learn, which may make it non-beneficial for this deployment, but such a system would be beneficial in future deployments where review boards may make the volunteer process a bit more difficult. The SCALE server could easily ignore data from these clients for the purposes of analytics and alerting.

Metadata

Metadata

Assignees

No one assigned

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions