Large-Scale Real-Time Product Recommendation At Criteo
by Romain Lerallut (Criteo)
Performance retargeting consists of displaying online advertisements that are personalized according to each user’s browsing history. Criteo’s recommender system chooses a dozen relevant products from over two billion candidates in a few milliseconds, not only for their click performance but also for their probability to generate a sale. In this talk, we will expose how to build such a system through a combination of offline and real-time computations, and the challenges of evolving it when regular A/B testing no longer suffices.
About the Speaker
Romain Lerallut is a Senior Engineering Manager at Criteo in the Engine department, in charge of applying large-scale machine learning algorithms to actual problems such as product recommendation or graphical layout optimization. Romain has been working on building up the Engine since 2011, first as devlead and then as manager. Before joining Criteo, he was teaching computers how to read cursive handwriting at A2iA. He has an engineering degree from “Ecole des Ponts-Paristech” and a PhD in Computer Science from “Ecoles des Mines-Paristech”.
Slides
http://www.slideshare.net/RomainLerallut/recsys-2015-largescale-realtime-product-recommendation-at-criteo
Challenges Encountered Scaling Up Recommendation Services
by Bottyán Németh (Gravity R&D)
Gravity R&D has been providing recommendation engines as SaaS solutions since 2009. The company has a strong research focus and recommendation quality has always been their primary differentiating factor. Widely used or open source recommendation algorithms are of little use to our technology team as a result of the superiority of our in-house developed, proprietary algorithms. Gravity R&D experienced many challenges while scaling up their services. The sheer quantity of data handled on a daily basis increased exponentially. This presentation will cover how overcoming these challenges permanently shaped our algorithms and system architecture used to generate these recommendations.
About the Speaker
Bottyán is one of the founders and has been working at Gravity R&D from the very beginning. He is directing the development of Gravity’s recommendation engine from several aspects. Bottyán oversaw the company scale up recommendations served per day from several million per day, to over 100 million per day. At the same time, he assisted Gravity in the addition of extra recommendation and personalization related features. He has contributed to this evolution in several ways with the development of the data mining tools, design of the system architecture, and development of the operation workflow. His current focus is on how to use state of the art recommendation algorithms to solve real world problems.
Slides
http://www.slideshare.net/domonkostikk/challenges-encountered-by-scaling-up-recommendation-services-at-gravity-rd
Recommendations in Travel
by Onno Zoeter (Booking.com)
Recommender systems have received much attention in recent years, and they have been successfully applied in many different domains. With each domain come new constraints that require system designers to make choices about how to apply and extend generic algorithms in their context. Booking.com is planet Earth’s number one accommodation reservation site. The accommodation recommendation problem that it needs to solve has several interesting and unique challenges; for example properties are not fixed like movies, but influenced by external, changing, factors like price, seasons and events. In this talk, we will highlight several challenges we have encountered and solutions we have developed.
About the Speaker
Dr. Onno Zoeter is principal data scientist at Booking.com where he works on designing and improving prediction, collaborative filtering and ranking models. Before joining Booking.com, Dr. Zoeter led the research on transportation demand management at Xerox’s European Research Lab and worked on large scale click-through rate prediction problems at Microsoft Research in Cambridge. His research received numerous awards. Among them the OECD International Transport Forum Innovation Award and the International Parking Institute Award of Excellence. The large scale system for advertisement relevance prediction he co-developed won Microsoft’s adPredict competition and was adopted as part of the Bing search engine. Fortune selected him among their 20 Big Data All-Stars in 2014.
Making Meaningful Restaurant Recommendations At OpenTable
by Sudeep Das (OpenTable)
At OpenTable, recommendations play a key role in connecting diners with restaurants. The act of recommending a restaurant to a diner relies heavily on aligning everything we know about the restaurant with everything we can infer about the diner. Our methods go beyond using the diner-restaurant interaction history as the sole input — we use click and search data, the metadata of restaurants, as well as insights gleaned from reviews, together with any contextual information to make meaningful recommendations. In this talk, I will highlight the main aspects of our recommendation stack built with Scala using Apache Spark.
About the Speaker
Sudeep is a data scientist at OpenTable with a passion for turning data into meaningful insights and stories. Currently, he is using an arsenal of data science tools to derive insights from OpenTable’s extensive data trove. Sudeep’s current projects include natural language processing and topic analysis on an extensive review corpus to reveal salient features of restaurants and reviewers. He is also a key contributor in building a restaurant recommendation stack using matrix factorization, content-based and hybrid factorization machine-based approaches. Sudeep loves using Python-based tools like Pandas, Scikit-learn, Gensim, PySpark, Matplotlib, and mpld3, as well as Scala and Spark.
For most of his professional life Sudeep has been an astrophysicist. He holds a PhD is Astrophysics from Princeton University.
Slides
http://www.slideshare.net/SudeepDasPhD/recsys-2015-making-meaningful-restaurant-recommendations-at-opentable
The Role of User Location in Personalized Search and Recommendation
by Ido Guy (YAHOO! Labs)
With mobile devices, users no longer access the web from specific locations, but virtually from anywhere. How does this affect our ability to provide personalized information for users? In this talk, I will discuss the influence of location activity on users’ information needs and how a better understanding of these needs can help enhance web applications in which personalization plays a key role.
About the Speaker
Ido Guy is a Principal Research Engineer at Yahoo Labs, where he focuses on data science for Yahoo products in search, questions answering, and email. Prior to joining Yahoo Labs, Ido was a senior technical staff member and manager of the Social Technologies group, an area that he established and managed in 2007-2014, as part of the IBM Haifa Research Lab. Recommender systems are a key domain of Ido’s research, with focus on the social media domain. He served as program co-chair for RecSys 2012.
