A Survey on Transfer Learning for Cross-Domain Recommendation

PhD Qualifying Examination


Title: "A Survey on Transfer Learning for Cross-Domain Recommendation"

by

Mr. Guangneng HU


Abstract:

Recommender systems (RSs) assist consumers in tackling the information overload 
and long-tail issues among millions of products and services. Collaborative 
filtering (CF) is the key technique for RSs. CF exploits user-item behavior 
interactions (e.g., clicks) only and suffers from the data sparsity issue. The 
cross-domain (CD) recommendation technique is an effective way of alleviating 
the data sparse issue in RSs by leveraging the knowledge from relevant domains. 
This matches the core idea of transfer learning (TL) which leverages knowledge 
from a source domain to improve predictive performance in a target domain that 
has no sufficient labeled data. This survey introduces transferring knowledge 
approaches for CD recommendation, including feature/parameter based transfer, 
instance relationship transfer, and pattern/model based transfer.

With the success of deep learning (DL) for learning high-level representations, 
there is a tendency towards applying DL for CD RSs, leading to the deep 
transfer learning for cross-domain recommendation. This survey also introduces 
these research frontiers and points out some promising directions for further 
investigation.


Date:			Thursday, 28 June 2018

Time:                  	5:30pm - 7:30pm

Venue:                  Room 5560
                         Lifts 27/28

Committee Members:	Prof. Qiang Yang (Supervisor)
 			Prof. Huamin Qu (Chairperson)
 			Dr. Yangqiu Song
 			Dr. Jianfeng Cai (MATH)


**** ALL are Welcome ****