A Survey on Session-Based Recommendation

PhD Qualifying Examination


Title: "A Survey on Session-Based Recommendation"

by

Mr. Weile TAN


Abstract:

Session-based recommendation (SBR) has emerged as a crucial paradigm in 
modern recommender systems, designed to predict a user's next action based 
solely on their short-term, ongoing interactions. Unlike sequential 
recommendation, SBR operates without long-term user profiles, making it 
highly applicable in scenarios with anonymous users or strict privacy 
constraints. In this survey, we provide a comprehensive review of the 
state-of-the-art in session-based recommendation. We systematically 
categorize existing approaches into classic item ID-based methods, spanning 
conventional techniques to modern deep learning architectures like sequential 
and graph-based models, and side information-driven methods that leverage 
temporal, attribute, behavioral, and multi-modal data. Furthermore, we 
summarize the standard evaluation protocols, including widely used datasets 
and metrics for both accuracy and diversity. Finally, we highlight the 
current challenges in the field and outline promising future directions, 
including the exploration of diverse and uncertain user intent 
representations, and the integration of emerging technologies such as 
diffusion models and large language models (LLMs). This survey aims to serve 
as a thorough reference for researchers and practitioners, facilitating 
further advancements in the dynamic field of session-based recommendation.


Date:                   Thursday, 23 April 2026

Time:                   10:00am - 12:00noon

Venue:                  Room 2132C
                        Lift 22

Committee Members:      Prof. Raymond Wong (Supervisor)
                        Prof. Dimitris Papadias (Chairperson)
                        Dr. Xiaojuan Ma