Recsperts - Recommender Systems Experts

In this first interview we talk to Kim Falk, Senior Data Scientist, multiple RecSys Industry Chair and author of the book "Practical Recommender Systems". We introduce into recommenders from a practical perspective discussing the fundamental difference between content-based and collaborative filtering as well as the cold-start problem - no mathematical deep-dive yet, but expect it to follow. In addition, we reason what constitutes good recommendations and briefly touch on a couple of ways of finding that out.Looking a bit into the history of the recommender systems community, we touch on the Netflix Prize that was running from 2006 to 2009 as well as on the RecSys - the leading conference in recommender systems, where we also met for the first time.In the end, we discuss a couple of challenges the field faces, in particular associated with approaches based on deep learning. Besides that, Spiderman will accompany our conversation at certain times. Plus many practical recommendations included on how to get started. Stay tuned!Links from this Episode:Kim Falk on LinkedIn and TwitterBook: Practical Recommender Systems (Manning) (get 37% discount with the code podrecsperts37 during checkout)GitHub Repository for PRS BookACM Conference on Recommender Systems 2021 (Amsterdam)Recommender Systems Specialization at CourseraAmazon.com Recommendations: Item-to-Item Collaborative FilteringNetflix PrizeNetflix Prize dataset on KaggleNew York Times: A $1 Million Research Bargain for Netflix, and Maybe a Model for OthersEvaluation Measures for Information RetrievalPaper by Dacrema et al. (2019): Are We Really Making Much Progress? A Worrying Analysis of Recent Neural Recommendation Approaches (best paper award at RecSys 2019)Recommending music on Spotify with Deep LearningMovieLens RecommendersGeneral Links:Follow me on Twitter: https://twitter.com/LivesInAnalogiaSend me your comments, questions and suggestions to marcel@recsperts.comPodcast Website: https://www.recsperts.com/Twitter and LinkedIn posts for sharing:LinkedInTwitter

Have you ever though about how Spotify is able to generate its fantastic Discover Weekly Playlist, how Amazon is generating a fortune by showing what other like you purchased in the past, or how Netflix achieves high user retention? The answer is personalization and in this show we focus on the most prominent way to achieve personalization: recommender systems.Whether you are a beginner and new to the field or you have already build recommenders, this show is to bring you the experts in recommender systems to share their knowledge and expertise with all of us. It is for making the topic more accessible and to provide a regular coverage of basics and advances in recommender systems research and application. I invite the experts to share their insights and to provide you with the right knowledge to get started and gain expertise yourself.In this introductory episode I am going to share some exemplary use cases from different industries (music streaming, e-commerce, travel, or social networks) along with challenges and problems in research and application. Plus, I am presenting the first guest for our upcoming episode.Links from the show:ACM Conference on Recommender Systems 2021 (Amsterdam): https://recsys.acm.org/recsys21/Introductory Python RecSys Training: https://github.com/mkurovski/recsys_trainingFollow me on Twitter: https://twitter.com/LivesInAnalogiaRead my RecSys Blogposts: https://medium.com/@marcel.kurovskiSend me your comments, questions and suggestions to marcel@recsperts.comPodcast Website: https://www.recsperts.com/