Privacy Preserving Similarity Evaluation of Time Series Data

Speaker:        Dr. George Kollios
                Boston University

Title:          "Privacy Preserving Similarity Evaluation of Time
                 Series Data"

Date:           Monday, 17 March 2014

Time:           4:00pm - 5:00pm

Venue:          Lecture Theater F (near lifts 25/26), HKUST

Abstract:

In this talk, I will discuss some recent results on privacy preserving
evaluation of the similarity between two time series that belong to two
different parties. The goal is the compute the similarity between the two
time series without revealing each time series to the other party. We use
the Dynamic Time Warping distance to define the similarity between two
time series and we present a protocol that tries to hide both the original
time series and the dynamic programming matrix that is used to compute the
similarity. In addition, we need to hide the path in the matrix that gives
the optimal solution. The protocol combines partial homomorphic encryption
and random offsets. However, our protocol, although it is orders of
magnitude faster than other existing methods, leaks some information and I
will discuss an idea to describe and quantify this leakage. An
experimental evaluation on some real datasets show that the proposed
approach is very promising.

In addition, if time permits, I will discuss some recent results on a
database-friendly encryption scheme for range queries.


********************
Biography:

George Kollios is an Associate Professor in the Computer Science
Department at Boston University in Boston, Massachusetts. He received his
Diploma in Electrical and Computer Engineering in 1995 from the National
Technical University of Athens, Greece; and the M.Sc. and Ph.D. degree in
Computer Science from Polytechnic University (now NYU-Poly), New York in
1998 and 2000 respectively. His research interests include temporal and
spatio-temporal indexing, data mining, database security, multimedia
indexing, and approximation algorithms for large-scale data management
problems. His research has been supported by NSF, including an NSF CAREER
Award, and IARPA. He is a member of ACM and IEEE Computer Society.