Multi-Modal Recommender Systems: Hands-On Exploration
Tuan, Aghiles, and Hady will be delivering a tutorial at the RecSys-21 conference that will take place in September 2021.
Recommender systems typically learn from user-item preference data such as ratings and clicks. This information is sparse in nature, i.e., observed user-item preferences often represent less than 5% of possible interactions. One promising direction to alleviate data sparsity is to leverage auxiliary information that may encode additional clues on how users consume items. Examples of such data (referred to as modalities) are social networks, item’s descriptive text, product images. The objective of this tutorial is to offer a comprehensive review of recent advances to represent, transform and incorporate the different modalities into recommendation models. Moreover, through practical hands-on sessions, we consider cross model/modality comparisons to investigate the importance of different methods and modalities. The hands-on will be conducted with Cornac, a comparative framework for multimodal recommender systems.
- Brief overview of recommender systems (20 minutes)
- Introduction to multimodal recommender systems (20 minutes)
- Hands-on: Starting with the Cornac framework (10 minutes)
- Exploration into each modality (90 minutes):
* Text modality
* Image modality
* Network modality
- Cross-modal utilization (30 minutes)
- Future directions (10 minutes)
Introductory to intermediate. We target both practitioners seeking applicable experience, as well as researchers interested in recent and future research directions in multimodal recommender systems.
Basic knowledge of Python, machine learning and recommender systems.
Quoc-Tuan Truong is a PhD student at Singapore Management University (SMU). His research focuses on multimodal representation learning and preference modeling for recommender systems. His teaching experiences include a tutorial on “Recommender Systems” at AI Singapore Summer School 2020, and a tutorial on “Facial Expression Recognition using CNN” organized by KDD.SG. More information can be found at his homepage.
Aghiles Salah is an Applied Research Scientist at the Rakuten Institute of Technology. His research is in machine learning and recommender systems. He is the (co) author of several publications in tier-1 conferences and journals, and he taught (over more than four years) several undergraduate and master’s courses on machine learning, data analysis, and related topics. More information may be found at his homepage.
Hady W. Lauw is an Associate Professor at SMU School of Computing and Information Systems and the current Chair of the Singapore Chapter of ACM SIGKDD (KDD.SG). He publishes actively on AI and recommender systems, earning a Distinguished Paper Award at IJCAI-20 and an Outstanding Paper Nomination at AAAI-14. He has also conducted tutorials in major conferences, such as AAAI-19, IJCAI-11, and CIKM-10. He has 10 years of university teaching experience (twice nominated for teaching awards) at both undergraduate and postgraduate levels, including a master-level course on recommender systems. More information can be found at his homepage.