Publication Type
PhD Dissertation
Version
publishedVersion
Publication Date
7-2025
Abstract
Traditional research in recommendation systems has largely centered on the static offline supervised learning setting. In this paradigm, all available user-item interaction data is collected and partitioned into fixed training, validation, and test sets. Models are developed and evaluated in this controlled environment, where the underlying data distribution is assumed to remain unchanged. This approach offers clear advantages: it simplifies experimentation, enables reproducible benchmarking, and allows for straightforward comparisons between algorithms.
However, this static offline setting does not reflect the realities faced by modern recommendation systems. In real-world applications, data is dynamic and ever-evolving, where new users and items are constantly emerged, user preferences shift over time, and interactions arrive as a continuous stream. Moreover, data is often fragmented across multiple platforms or domains, each with its own characteristics and challenges. These factors introduce complexities such as the need for continual adaptation and transferring knowledge across domains.
Recognizing these limitations, this dissertation aims to bridge the gap by formulating recommendation problems that more accurately reflect real-world scenarios. The primary goal is to design dynamic frameworks that efficiently learn from new data streams, capably handle evolving user behaviors and item catalogs, and effectively address data fragmentation across platforms by enabling knowledge transfer between various tasks and domains.
Degree Awarded
PhD in Computer Science
Discipline
Computer and Systems Architecture | Systems Architecture
Supervisor(s)
LAUW, Hady Wirawan
First Page
1
Last Page
114
Publisher
Singapore Management University
City or Country
Singapore
Citation
DO, Dinh Hieu.
Modeling multiple tasks in recommendation systems. (2025). 1-114.
Available at: https://ink.library.smu.edu.sg/etd_coll/792
Copyright Owner and License
Author
Creative Commons License
This work is licensed under a Creative Commons Attribution-NonCommercial-No Derivative Works 4.0 International License.