A Survey On Transfer Learning in Urban Computing

PhD Qualifying Examination


Title: "A Survey On Transfer Learning in Urban Computing"

by

Miss Yexin LI


Abstract:

Recently, rapid urbanization progress has modernized and facilitated 
people’s lives significantly while engendering some severe problems, e.g. 
traffic congestion, environment pollution, energy consumption, etc. Urban 
computing, as an interdisciplinary research field, tries to use multiple 
heterogeneous data sources generated in human’s daily activities to solve 
these problems from a big data perspective. However, not every city has 
the advanced technologies and infrastructures to measure and store these 
heterogeneous data sources, making the current urban computing models 
perform poorly sometimes, i.e. in the data-sparse cities. As another 
famous research area in these decades, transfer learning, which can 
transfer knowledge from a data-rich source domain to a data-sparse target 
domain, can address the data-sparsity issue excellently. Therefore, it is 
intuitive for us to consider adopting transfer learning to urban computing 
to develop more general models that can not only perform well in cities 
which have enough heterogeneous data but also in those data-sparse ones.

This article firstly introduces urban computing definition, its frequently 
used data sources and some representative urban problems, which are 
categorized into four groups. After that, we summarize three categories of 
representative transfer learning methods to illustrate how knowledge is 
usually transferred between different domains. Thirdly, we combine 
transfer learning and urban computing by introducing some representative 
works, which adopt transfer learning to solve urban problems better than 
the traditional ones. We conclude by discussing some unsolved issues met 
in traditional urban computing, which may be addressed by transfer 
learning, e.g. the model training efficiency issue in urban planning.


Date:			Tuesday, 12 June 2018

Time:                  	4:00pm - 6:00pm

Venue:                  Room 5560
                         Lifts 27/28

Committee Members:	Prof. Qiang Yang (Supervisor)
 			Prof. Dit-Yan Yeung (Chairperson)
 			Dr. Xiaojuan Ma
 			Dr. Yangqiu Song


**** ALL are Welcome ****