Taming Latent Factor Models for Explainability
Assistant Professor, Department of Computer Science, University of Virginia
Time: November 6, 2019 @ 11:15 AM to 12:05 PM
Location: 204 Evans Hall
Latent factor models have achieved great success in personalized recommendations, but they are also notoriously difficult to explain. In this talk, I will share our recent effort in integrating regression trees to guide the learning of latent factor models for explainable recommendation. Specifically, we build regression trees on users and items recursively based on user-provided review content, and associate a latent factor to each node on the trees to represent users and items for recommendation. With the growth of regression tree, we are able to track the creation of latent factors by looking into the path of each factor on regression trees, which thus serves as an explanation for the resulting recommendations. If time allows, I would also like to share our progress in multi-task tensor factorization and generative neural collaborative filtering for explainable recommendation.