Monday, December 19, 2022 - 13:00 to 14:30
SCHOOL OF COMPUTER SCIENCE
The School of Computer Science is pleased to present…
MSc Thesis Proposal by: Lama Khalil
Date: Monday, December 19, 2022
Time: 1:00 pm – 2:30 pm
Location: Essex Hall, Room 122
Advances in joint recommendation and knowledge graph completion (KGC) learning have enhanced the precision and explainability of recommendations. Recent studies have established that consideration of the incomplete nature of knowledge graphs (KGs) improves the performance of recommender systems (RSes). However, the existing multi-task learning models (MTLs) rely on translation-based knowledge graph embedding (KGE) techniques for KGC that fail to capture different relation patterns, such as composition relations. This type of relation is prevalent in real-world KGs. In this study, we propose a simple but effective approach to enhance the KGC task while training it with the RS. Our approach, rotational knowledge-enhanced translation-based user preference (RKTUP), is an advanced variant of the knowledge-enhanced translation-based user preference model (KTUP), an existing MTL model. To enhance KTUP, we use the rotational-based KGE technique (RotatE) to model and infer various relation patterns, such as symmetry/asymmetry, composition, and inversion. Unlike earlier MTL models, RKTUP can model and infer diverse relation patterns while learning more robust representations of the entities and relations in the KGC task to recommend the top-N items for a user. The experimental results reveal that RKTUP outperforms existing methods and achieves state-of-the-art performance on the KGC and recommendation tasks.
Keywords: Recommender System, Knowledge Graph Completion, Multi-Task Learning Model.
MSc Thesis Committee:
Internal Reader: Dr. Kalyani Selvarajah
External Reader: Dr. Mohammad Hassanzadeh
Advisor: Dr. Ziad Kobti
MSc Thesis Proposal Announcement
5113 Lambton Tower 401 Sunset Ave. Windsor ON, N9B 3P4 (519) 253-3000 Ext. 3716 firstname.lastname@example.org