The tutorials will take place in BC 01 on the 23rd of October.
Tutorial 1: Robust Recommender Systems
Robin Burke (DePaul University)
Recommender systems bring thousands and sometimes millions of users together to share opinions, automating the large-scale sharing of expertise. However, some of these users are not genuine enthusiasts seeking to share their experiences, but rather marketers, spammers and vandals, all of whom seek to bias a system's results. It is obvious that a sufficiently large population of such malicious users will render any collaborative system useless. What options do system implementers and maintainers have? The purpose of this tutorial is to explore some of the recent research in the area of recommender system robustness. In particular, we will look at recent work in attack modeling, the quantification of attack impact, alternate recommendation algorithms, trust maintenance, and attack detection.
Tutorial 2: Recent Progress in Collaborative Filtering
Yehuda Koren (Bell Labs)
The Netflix Prize competition brought a lot of excitement and new researchers to the Collaborative Filtering field. Naturally, the surge of interest led to many innovative ideas and improved algorithms. In this tutorial I will survey Collaborative Filtering techniques, while demonstrating their use on the Netflix data. Description will include some basic techniques together with the most recent developments in the field, which played central role in advancing the analysis of the Netflix dataset. Recommender systems strive to achieve a multidimensional set of goals. Hence, multiple evaluation metrics are needed in order to assess recommendation quality. Some of these measures will be described and compared in the tutorial. In addition, I will emphasize the possible integration of extra data attributes such as content (actors, directors, genre, etc.), implicit feedback and rating dates.
Tutorial 3: Context-Aware Recommender Systems
Gedas Adomavicius (University of Minnesota) and Alex Tuzhilin (NYU)
Traditionally recommender systems have been focusing on recommending the most relevant items to users or the most appropriate users to items. While the traditional recommendation technologies have performed reasonably well in several applications, in many other applications, such as location- and time-based services, including travel recommendations, it may not be sufficient to consider only users and items - it is also important to incorporate contextual information into the recommendation process. In this tutorial, we will review various ways of providing the contextual information and incorporating it into recommendation algorithms and suggest possible research directions in this area of recommender systems.