NOVELTY AND DIVERSITY BASED RECOMMENDATION SYSTEMS

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


PhD Thesis Defence


Title: "NOVELTY AND DIVERSITY BASED RECOMMENDATION SYSTEMS"

By

Mr. Pengfei ZHAO


Abstract

Traditional recommendation systems aim at generating recommendations that 
are relevant to the user's interest which are named as relevance-based 
recommendation systems (RBRS). The major drawback of this approach is that 
the user soon becomes very familiar with the recommendations and loses 
interest in reading and exploring them. Discovery-oriented recommendation 
systems (DORS) aim to solve this problem by introducing discover utilities 
(DU's) such as novelty, diversity and serendipity to improve the 
attractiveness of the recommendations to the user. In this thesis, we 
investigate techniques for improving the effectiveness of DORS, 
specifically on the perspective of novelty and diversity, which are most 
important and widely studied DU's.

Existing DORS generates recommendations that are optimized to balance 
between the accuracy and DU's of the recommendations to make the 
recommendations relevant and yet interesting to the user. However, they 
disregard an important fact that different users' appetites for DU's are 
different. For example, a curious user can accept highly novel and 
diversified recommendations but a conservative user tends to respond only 
to recommendations she is familiar with. We propose a framework for 
curiosity-based recommendation system (CBRS) which can produce 
recommendations with an amount of DU's personalized to fit individual 
user's curiosity level. As a result, the recommendations are neither too 
surprising nor too boring for a user because the recommendations are 
customized to fit her unique curiosity. In order to model and quantify 
human curiosity, we adopt the curiosity arousing model (CAM) developed in 
psychology research and propose a probabilistic curiosity model (PCM) to 
model the psychological model computationally.

To improve the diversity of the recommendations, we propose a 
recommendation framework by the unification of two types of diversities, 
namely, intra-list and temporal diversity, of the recommendations. 
Traditional RBRSs recommend items which are very similar to the user's 
interest. As a result, the recommended items are also very similar between 
each other, making the items in a recommendation list monotonous. We name 
this "intra-monotony problem" (IMP). Further, most existing recommendation 
systems make recommendations without considering what has been recommended 
before. Thus, they may make similar recommendations over and over again, 
making the recommended items across recommendation lists monotonous. We 
name this "temporal monotony problem" (TMP). To address the two problems, 
previous research has utilized intra-list diversity (intraD) and temporal 
diversity (timeD) to improve, respectively, the diversity within a 
recommendation list and across recommendation lists. However, existing 
work studies these two diversity types separately. We propose an approach 
to unify the two diversity types into a single framework so that both 
intra-list and temporal diversity can be considered holistically. Rather 
than arbitrarily combining intraD and timeD, we propose a new diversity 
type called jointD and optimize it by formulating the problem as a 
constraint quadratic optimization problem. This approach allows both 
intraD and timeD to be jointly processed.


Date:			Thursday, 18 August 2016

Time:			2:00pm – 4:00pm

Venue:			Room 3494
 			Lifts 25/26

Chairman:		Prof. Xun Wu (SOSC)

Committee Members:	Prof. Dik-Lun Lee (Supervisor)
 			Prof. Ke Yi
 			Prof. Nevin Zhang
 			Prof. Tat-Koon Koh (ISOM)
 			Prof. Kam-Fai Wong (Sys Engg & Engg Mgmt, CUHK)


**** ALL are Welcome ****