Privacy for Recommender Systems
by Bart Knijnenburg (Clemson University, USA) and Shlomo Berkovsky (CSIRO, Australia)
Websites increasingly gather tremendous amounts of user data for recommendation purposes. This data may pose a severe threat to user privacy, e.g., if accessed by untrusted parties, or used inappropriately. Hence, it important for recommender system designers and service providers to learn about ways to generate accurate recommendations while at the same time respecting the privacy of their users. In this tutorial, we will:
- analyze common privacy risks imposed by recommender systems
- survey architectural, algorithmic, policy-related, and UI-design solutions
- discuss implications for users
This tutorial is of general interest and is relevant for both participants with longstanding experience in recommender systems, as well as to newcomers. No specific background or skills are required. The tutorial will conclude with a plenary discussion of the future of privacy in recommender systems.
Slides
https://www.dropbox.com/s/55sdr62pqmpr20p/privacy-tutorial.pdf
Date
Monday, Aug 28, 2017, 16:15-18:00
Location
Main Room
New Paths in Music Recommender Systems Research
by Markus Schedl (Johannes Kepler University Linz, Austria), Peter Knees (Vienna University of Technology, Austria) and Fabien Gouyon (Pandora Inc., USA)
In the RecSys community, music is too often treated as “just another item”. Yet, the particularities of music data and its multiple modalities open many opportunities, e.g., to leverage content-based audio features or to build comprehensive listener models that go beyond simple user-item interactions. Furthermore, since it is now increasingly more common for a music listener to simply stream music rather than to purchase and own it, today’s music recommenders need to focus on recommending a listening experience. Algorithms that produce a one-shot recommendation for the purpose of a track or album purchase are no longer of central importance. As a consequence, Music Recommender System (MRS) research has to face a wide range of challenges, such as sequential recommendation, or conversational and contextual recommendation.
This introductory tutorial incorporates both academic and industrial points of view on latest developments in music recommendation research, presenting challenges and solutions. The content will be organized with respect to three use cases: playlist generation, context-aware music recommendation, and recommendation in the creative process of music making. In addition, we will discuss the implications of recent MRS technologies on actors, other than the listener, in the rich and complex music industry ecosystem (e.g., labels, music makers and producers, concert halls, advertisers). No particular prerequisite knowledge or skills are required from the audience, other than a very basic understanding of the main concepts in recommender systems. Accompanying the tutorial, we will publish a comprehensive set of slides, including references to state-of-the-art work and open implementations of several of the presented techniques.
Slides
http://www.cp.jku.at/tutorials/mrs_recsys_2017/
Date
Monday, Aug 28, 2017, 16:15-18:00
Location
Room 1
Deep Learning for Recommender Systems
by Alexandros Karatzoglou (Telefonica Research, Spain) and Balázs Hidasi (Gravity R&D, Hungary)
The past few years have seen the tremendous success of deep neural networks in a number of complex machine learning tasks such as computer vision, natural language processing and speech recognition. For these reasons, Deep Learning has been hailed as the “next big thing” in recommender systems, and we have started to see deep neural networks deliver on their potential for dramatic improvement in Recommendation Systems technology.
The aim of the tutorial is dual, 1) to introduce deep learning techniques that have been and are used in recommender systems such as Recurrent Neural Networks and Convolutional Networks 2) to present the current state-of-the-art collaborative filtering and content-based methods that use deep learning techniques to provide recommendations. The tutorial does not require any prior knowledge in Deep Learning since there will be detailed introductions to the relevant techniques, e.g., Recurrent Neural Networks, Convolutional Networks, word2vec embeddings.
Slides
https://www.slideshare.net/…/deep-learning-for-recommender-systems-recsys2017-tutorial
Date
Tuesday, Aug 29, 2017, 16:15-18:00
Location
Main Room
Product Recommendations Enhanced with Reviews
by Muthusamy Chelliah (Flipkart, India) and Sudeshna Sarkar (IIT Kharagpur, India)
E-commerce websites commonly deploy recommender systems that make use of user activity (e.g., ratings, views, and purchases) or content (product descriptions). These recommender systems can benefit enormously by also exploiting the information contained in customer reviews. Reviews capture the experience of multiple customers with diverse preferences, often on the fine-grained level of specific features of products. Reviews can also identify consumers’ preferences for product features and provide helpful explanations. The usefulness of reviews is evidenced by the prevalence of their use by customers to support shopping decisions online. With the appropriate techniques, recommender systems can benefit directly from user reviews.
This tutorial will present a range of techniques that allow recommender systems in e-commerce websites to take full advantage of reviews. Topics covered include text mining methods for feature-specific sentiment analysis of products, topic models and distributed representations that bridge the vocabulary gap between user reviews and product descriptions, and recommender algorithms that use review information to address the cold-start problem.
The tutorial sessions will be interspersed with examples from an online marketplace (i.e., Flipkart) and experience with using data mining and Natural Language Processing techniques (e.g., matrix factorization, LDA, word embeddings) from Web-scale systems.
Slides
https://www.slideshare.net/maranlar/product-recommendations-enhanced-with-reviews
Date
Tuesday, Aug 29, 2017, 16:15-18:00
Location
Room 1
Open Source Tools for Online Learning Recommenders
by Róbert Pálovics (Hungarian Academy of Sciences, Hungary), Domokos Kelen (Hungarian Academy of Sciences, Hungary) and András A. Benczúr (Hungarian Academy of Sciences, Hungary)
Recommender systems have to serve in online environments that can be non-stationary.
Traditional recommender algorithms may periodically rebuild their models, but they cannot adjust to quick changes in trends caused by timely information. In contrast, online learning models can adapt to temporal effects, hence they may overcome the effect of concept drift.
As a new experiment at RecSys, we provide a hands-on tutorial to present open source systems capable of updating their models on the fly after each event: Apache Spark, Apache Flink and Alpenglow, our new release C++ recommender system with Python API.
Participants of the tutorial will be able to experiment with all the three systems by using interactive Zeppelin and Jupyter Notebooks on their own laptops.
The final objective of the tutorial is to compare and then blend batch and online methods to build models providing high quality top-k recommendation in non-stationary environments. Participants should bring their own laptops and prepare for a hands-on tutorial. Required theoretical background will be summarized and published before the conference.
Pre-installing the required software is recommended, but not mandatory: organizers will help the audience with the installation steps at the tutorial.
Materials
https://github.com/rpalovics/recsys-2017-online-learning-tutorial
Date
Hands-on tutorial running parallel to the workshops at Sunday, Aug 27, 2017.
Participants may attend the tutorial in two identical sessions starting from either 9:00 or 14:00.
Both sessions start with installation instructions, which is crucial to participate.
Location
Room 7
