Platforms like Amazon and YouTube train their recommender systems by understanding user interest from their interaction history. In the normal days, no problem. But during extraordinary events like the COVID-19 pandemic, people tend to exhibit unusual behavior, such as panic buying. My work is to differentiate between genuine user interests and irrational behavior, in order to improve the quality of recommendations.
PhD in Recommender Systems, 2020
Master of Information Technology, 2018
Bachelor of Mechanical Engineering, 2010
Beijing Information Science & Technology University
CNN, RNN, GRU, Transformer, GNN, Reinforcement Learning, etc.
Python, R, etc.
React, Vue, Node.js, Laravel, Sails JS, Asp.NET, etc.
Android, iOS, etc.
My lovely little cats: Kiki (奇奇) and MiaoMiao (妙妙).
Hiking in the mountains, enjoying the beauty of nature.
Code for a Better World: Programming Openly, Transforming Ideas into Impact.
Switch, PC, etc.
We outline a simulation-based study of the effect rapid population-scale concept drifts have on Collaborative Filtering (CF) models. We create a framework for analyzing the effects of macro-trends in population dynamics on the behavior of such models. Our framework characterizes population-scale concept drifts in item preferences and provides a lens to understand the influence events, such as a pandemic, have on CF models. Our experimental results show the initial impact on CF performance at the initial stage of such events, followed by an aggravated population herding effect during the event. The herding introduces a popularity bias that may benefit affected users, but which comes at the expense of a normal user experience. We propose an adaptive ensemble method that can effectively apply optimal algorithms to cope with the change brought about by different stages of the event.