Collaborative and Transfer Learning in Recommendations

The Hong Kong University of Science and Technology
Department of Computer Science and Engineering


PhD Thesis Defence


Title: "Collaborative and Transfer Learning in Recommendations"

By

Mr. Bin CAO


Abstract

Nowadays recommendation has become a key feature in many online services. 
The quality of recommendations is one of the key factors to the revenue 
for these service providers. Therefore, it is critical for them to provide 
high quality recommendations.

However, the recommendation task is non-trivial. Learning problems in 
recommendation usually are not well-defined traditional learning problems. 
It is common that users would provide heterogeneous feedback on items from 
heterogeneous domains. For example, as a recommender system like Amazon, 
it may need to provide personalized movie recommendations for a user based 
her/his feedback from multiple domains including books and clothing. As a 
search engine like Google, it may need to serve personalized ads based on 
users' searching behaviors and browsing history.

In this thesis, we consider the problems of using transfer-learning 
techniques to improve recommendations. More specifically, we consider the 
problems where we formulate multiple recommendation tasks in the problem 
and we ask three questions. Firstly, how to share knowledge when items 
span over multiple heterogeneous domains? Secondly, how to share knowledge 
across different user feedback? Thirdly, how to transfer knowledge with 
respect to a particular recommendation task from other source tasks? To 
answer these questions, we study transfer-learning based collaborative 
filtering models that could handle heterogeneous feedback and domain 
adaptation. Furthermore, we conduct experiments on real world dataset to 
show the effectiveness of the proposed models.


Date:			Wednesday, 3 August 2011

Time:			10:00am – 12:00noon

Venue:			Room 3584
 			Lifts 27/28

Chairman:		Prof. David Hui (CBME)

Committee Members:	Prof. Qiang Yang (Supervisor)
 			Prof. Raymond Wong
 			Prof. Dit-Yan Yeung
                         Prof. Weichuan Yu (ECE)
                         Prof. Michael R. Lyu (Comp. Sci. & Engg., CUHK)


**** ALL are Welcome ****