Campus Event Calendar
Event Entry
What and Who
Accurate Analysis of Large Private Datasets
Vibhor Rastogi
University of Washington
SWS Colloquium
Vibhor Rastogi is a doctoral candidate in the Database group at the
University of Washington. His dissertation develops techniques for
privacy-preserving data analysis. His other research interests include
data uncertainty, data cleaning, and problems in large-scale data
management.
SWS, RG1
AG Audience
English
Note: We use this to send email in the morning.
Date, Time and Location
Thursday, 15 April 2010
10:00
90 Minutes
E1 5
5th floor
Saarbrücken
Abstract
Today, no individual has full control over access to his personal
information. Private data collected by various hospitals and
universities, and also by websites like Google and Facebook, contain
valuable statistical facts that can be mined for research and analysis,
e.g., analyze outbreak of diseases, detect traffic patterns on the road,
or understand browsing trends on the web, but concerns about individual
privacy severely restricts its use, e.g., privacy attacks led AOL to
recently pull-off its published search-log data.
To remedy this, much recent work focuses on data analysis with formal
privacy guarantees. This has given rise to differential privacy
considered by many as the golden standard of privacy. However, few
practical techniques satisfying differential privacy exist for complex
analysis tasks (e.g., analysis involving complex query operators), or
new data models (e.g., data having temporal correlations or
uncertainty). In this talk, I will discuss techniques that fill this
void. I will first discuss a query answering algorithm that can handle
joins (previously, no private technique could accurately answer join
queries arising in many analysis tasks). This algorithm makes several
privacy-preserving analyses over social network graphs possible for the
first time. Then I will discuss a query-answering technique over
time-series data, which enables private analysis of GPS traces and other
temporally-correlated data. Third, I will discuss an access control
mechanism for uncertain data, which enables enforcing security policies
on RFID-based location data.
Finally, I will conclude by discussing some privacy and security
problems in building next-generation computing systems based on new
models for data (e.g., uncertain data), computing (e.g., cloud
computing), and human computer interaction (e.g., ubiquitous systems).
information. Private data collected by various hospitals and
universities, and also by websites like Google and Facebook, contain
valuable statistical facts that can be mined for research and analysis,
e.g., analyze outbreak of diseases, detect traffic patterns on the road,
or understand browsing trends on the web, but concerns about individual
privacy severely restricts its use, e.g., privacy attacks led AOL to
recently pull-off its published search-log data.
To remedy this, much recent work focuses on data analysis with formal
privacy guarantees. This has given rise to differential privacy
considered by many as the golden standard of privacy. However, few
practical techniques satisfying differential privacy exist for complex
analysis tasks (e.g., analysis involving complex query operators), or
new data models (e.g., data having temporal correlations or
uncertainty). In this talk, I will discuss techniques that fill this
void. I will first discuss a query answering algorithm that can handle
joins (previously, no private technique could accurately answer join
queries arising in many analysis tasks). This algorithm makes several
privacy-preserving analyses over social network graphs possible for the
first time. Then I will discuss a query-answering technique over
time-series data, which enables private analysis of GPS traces and other
temporally-correlated data. Third, I will discuss an access control
mechanism for uncertain data, which enables enforcing security policies
on RFID-based location data.
Finally, I will conclude by discussing some privacy and security
problems in building next-generation computing systems based on new
models for data (e.g., uncertain data), computing (e.g., cloud
computing), and human computer interaction (e.g., ubiquitous systems).
Contact
Bettina Bennett
0631-93039602
--email hidden
Video Broadcast
Yes
Kaiserslautern
G26
206
passcode not visible
logged in users only
Bettina Peden-Bennett, 04/12/2010 14:06 -- Created document.