Computer science colloquia: Ying Cai
| Date/Time: | Thursday, 23 Sep 2010 at 3:40 pm |
|---|---|
| Location: | B29 Atanasoff |
| Cost: | Free |
| Phone: | 515-294-6516 |
| Channel: | College of Liberal Arts and Sciences |
| Categories: | Lectures |
| Actions: | Download iCal/vCal | Email Reminder |
Applications like location-based services provide valuable information to users, but also present them heightened privacy risks. A user's whereabouts may imply sensitive information, such as health condition and lifestyle. Simply using a pseudonym is not sufficient for self-protection. This is due to the fact that the location itself may reveal a user's real-world identity. A single location sample may not be linked directly to a particular user, but the accumulation of a time-series sequence of her location samples will eventually do so. This talk concerns the problems of location privacy protection in the context of applications wherein users' location data is subject to risks such as potential misuse by insiders, unintentional or mistaken disclosure. We will introduce a location cloaking technique that can prevent anonymous location data, either a single location sample or a time-series sequence, from being correlated with restricted spaces such as home and office for subject re-identification. We will also present a new approach for privacy requirement modeling. Compared to the traditional anonymity-K model, this scheme provides a much more intuitive way for a user to identify a desired level of location privacy protection.