Security and Privacy

Mainstream authentication and authorization mechanisms have been shown to be awkward and obtrusive by the ubiquitous computing community.

In mixed reality/smart spaces it is often the case that multiple users freely share multiple devices in collaborative sessions, e.g. a medical team sharing several devices in a treatment room, or a family sharing the devices in a smart home: how should the system keep track of who is interacting with each device at a given time, without constantly asking users to enter their credentials?

I envision combining multi-factor unobtrusive solutions, such as face recognition, tracking the location of personal devices, etc. with context awareness and probabilistic inference techniques to identify users.

Accurate user identification is critical for authorization, however, user natural behaviors include borrowing each other’s credentials as a form of delegation.  I envision factoring in social networks to detect the real identity of a user, while inferring intended delegations.

See related publications.