Learning with Sparse Data in Mobile Computing

PhD Thesis Proposal Defence


Title: "Learning with Sparse Data in Mobile Computing"

by

Mr. Wenchen Zheng


ABSTRACT:

Human behavior understanding from sensor observations is a useful task in 
both artificial intelligence and mobile computing. It is also a difficult 
task as the sensor/behavior data are usually noisy and sparse. In this 
proposal, we study the data sparsity problem in three major categories of 
applications in mobile computing, including location estimation, activity 
recognition and mobile recommendation. In each category of problems, we 
show that most of the existing learning algorithms suffer from the data 
sparsity problem and thus propose some solution which is able to 
incorporate as much auxiliary data as possible to boost the performance. 
These solutions explore all the user behavior’s key components, including 
user, location, activity and time, thus giving us an interesting point of 
view on mobile computing.


Date:                   Wednesday, 20 April 2011

Time:                   10:00am - 12:00noon

Venue:                  Room 3402
                         lifts 17/18

Committee Members:      Prof. Qiang Yang (Supervisor)
                         Prof. Dik-Lun Lee (Chairperson)
 			Dr. Lei Chen
 	                Prof. Dit-Yan Yeung


**** ALL are Welcome ****