Sylvain Castagnos

Modeling Preferences in a Distributed Recommender System

A good way to help users finding relevant items on document platforms consists in suggesting content in accordance with their preferences. When implementing such a recommender system, the number of potential users and the confidential nature of some data should be taken into account. This paper introduces a new P2P recommender system which models individual preferences and exploits them through a user-centered filtering algorithm. The latter has been designed to deal with problems of scalability, reactivity, and privacy.

Personalized communities in a distributed recommender system

The amount of data exponentially increases in information systems and it becomes more and more difficult to extract the most relevant information within a very short time. Among others, collaborative filtering processes help users to find interesting items by modeling their preferences and by comparing them with users having the same tastes. Nevertheless, there are a lot of aspects to consider when implementing such a recommender system. The number of potential users and the confidential nature of some data are taken into account. This paper introduces a new distributed recommender system based on a user-based filtering algorithm. Our model has been transposed for Peer-to-Peer architectures. It has been especially designed to deal with problems of scalability and privacy. Moreover, it adapts its prediction computations to the density of the user neighborhood.

Toward a Robust Diversity-Based Model to Detect Changes of Context

Being able to automatically and quickly understand the user context during a session is a main issue for recommender systems. As a first step toward achieving that goal, we propose a model that observes in real time the diversity brought by each item relatively to a short sequence of consultations, corresponding to the recent user history. Our model has a complexity in constant time, and is generic since it can apply to any type of items within an online service (e.g. profiles, products, music tracks) and any application domain (e-commerce, social network, music streaming), as long as we have partial item descriptions. The observation of the diversity level over time allows us to detect implicit changes. In the long term, we plan to characterize the context, i.e. to find common features among a contiguous sub-sequence of items between two changes of context determined by our model. This will allow us to make context-aware and privacy-preserving recommendations, to explain them to users. As this is an on-going research, the first step consists here in studying the robustness of our model while detecting changes of context. In order to do so, we use a music corpus of 100 users and more than 210,000 consultations (number of songs played in the global history). We validate the relevancy of our detections by finding connections between changes of context and events, such as ends of session. Of course, these events are a subset of the possible changes of context, since there might be several contexts within a session. We altered the quality of our corpus in several manners, so as to test the performances of our model when confronted with sparsity and different types of items. The results show that our model is robust and constitutes a promising approach.

From Community Detection to Mentor Selection in Rating-Free Collaborative Filtering

The number of items that users can now access when navigating on the Web is so huge that these might feel lost. Recommender systems are a way to cope with this profusion of data by suggesting items that fit the users needs. One of the most popular techniques for recommender systems is the collaborative filtering approach that relies on the preferences of items expressed by users, usually under the form of ratings. In the absence of ratings, classical collaborative filtering techniques cannot be applied. Fortunately, the behavior of users, such as their consultations, can be collected. In this paper, we present a new approach to perform collaborative filtering when no rating is available but when user consultations are known. We propose to take inspiration from local community detection algorithms to form communities of users and deduce the set of mentors of a given user. We adapt one state-of-the-art algorithm so as to fit the characteristics of collaborative filtering. Experiments conducted show that the precision achieved is higher then the baseline that does not perform any mentor selection. In addition, our model almost offsets the absence of ratings by exploiting a reduced set of mentors.