Title: Recent Advance on Recommendation Methods for Implicit Feedback

Abstract: In recent years, the focus of recommender system research has shifted from explicit feedback problems such as rating prediction to implicit feedback problems. In this talk, I will introduce our recent two works for addressing the recommendation problem with implicit feedback.

In our first work [He et al. WWW 2017], we strive to develop techniques based on neural networks for recommendation. Although many recent work has employed deep learning for recommendation, they primarily used it to model auxiliary information, such as textual descriptions of items and acoustic features of music. When it comes to model the key factor in collaborative filtering (CF) --- the interaction between user and item features, they still resorted to matrix factorization and applied an inner product on the latent features of users and items. In this work, we explore the use of neural networks for modeling user-item interactions, developing a series of neural models for the CF task.

In our second work [Bayer and He et al. WWW 2017], we propose a novel and generic optimization method on par with BPR (Bayesian Personalized Ranking) for learning recommender models from implicit feedback. Our iCD (implicit Coordinate Descent) method optimizes the expensive regression loss that accounts for the prediction on the whole data, but with a very cheap time complexity that is determined by the size of observed data only. We illustrate this framework on a variety of complex recommender models including factorization machines, SVDfeature, Parallel Factor Analysis and Tucker Decomposition. Overall, this work provides the theory and building blocks to derive efficient implicit CD algorithms for (multi-)linear recommender models.

Dr. Xiangnan He is currently a postdoctoral research fellow with the Lab for Media Search, National University of Singapore. His research interests span recommender system, information retrieval, multi-media and natural language processing. His works have appeared in several top-tier conferences such as SIGIR, WWW, MM, CIKM and AAAI, and top-tier journals including TKDE and TOIS. His work on recommender system has received the Best Paper Award Honorable Mention of ACM SIGIR 2016. Moreover, he has served as the PC member for the prestigious conferences including SIGIR, WWW, MM, CIKM and EMNLP, and invited reviewer for prestigious journals including TKDE, WWWJ and TIIS.