Speakers: Dr Shoujin Wang, Dr Liang Hu

Authors: Dr Shoujin Wang, Dr Liang Hu, Prof. Yan Wang, Prof. Longbing Cao, Prof. Quan Z. Sheng, and Prof. Mehmet A. Orgun

Time: Jan 8, 2021

Location: Online event (Pacifico Convention Plaza, Yokohama, Japan)

Abstract:

Nowadays, the renaissance of artificial intelligence (AI) has attracted huge attention from every corner of the world. Specially, machine learning approaches have deeply involved in AI research in almost all areas, e.g., natural language processing (NLP), computer vision (CV) and planning. In particular, recommender systems (RS), as probably one of the most widely used AI systems, has integrated into every part of our daily life. In this AI age, on the one hand, state-of-the-art machine learning approaches, e.g. deep learning, have become the primary choice to build RS, on the other hand, both the theories and applications of RS are developing rapidly.

Classic RS, e.g., collaborative filtering and content-based filtering, are mainly conducted on the users’ feedback or items’ contents to promote the items by predicting the users’ explicit preference (e.g., ratings) over items in the e-commerce area. This not only downgrade the performance of RS by ignoring other relevant information, but also greatly limits the application scenarios and domains of RS. This actually motivates the necessity of new theories and approaches for building next-generation RS, as well as developing the advanced applications of RS. In practice, in addition to explicit preference prediction in the e-commerce area, RS are more and more widely used in many emerging scenarios (e.g., implicit ranking prediction over items) and domains (e.g., FashionAI) for smarter decisions in recent years. This not only broadens the application scope of RS, but also benefits us from nearly every aspects including eating, dressing, living and traveling. To this end, this tutorial will first introduce the latest advanced theories and approaches to built the next-generation RS and then demonstrate the advanced applications of RS in emerging cases and domains. Particularly, in this tutorial, first, we will present the background and foundations of RS, followed by the illustration of three typical theories and approaches for building various next-generation RS: (1) sequential or session-based RS, (2) graph learning based RS, and (3) interactive and conversational RS, together with their prototypes. Finally, we will demonstrate emerging real-world applied cases of RS in FashionAI, FinTech and healthcare.