A/B Testing: Innovation @ Internet Scale
by Ya Xu (LinkedIn)
A/B testing is at the core of future innovations for recommender systems. Building an A/B platform that properly sets up, manages experiments, and produces trustworthy results at large scale is challenging already. It is even more difficult to make sure that conclusions drawn are meaningful, reproducible and stable over time. In this talk, we will focus on several challenges we have encountered and solutions we have developed as LinkedIn embraces A/B testing as the driver for innovation.
Ya has been working in the domain of online A/B testing for over 4 years. She currently leads a team of engineers and data scientists building a world-class online A/B testing platform @ LinkedIn. She also spearheads taking LinkedIn’s A/B testing culture to the next level by evangelizing best practices and pushing for broad-based platform adoption. Before LinkedIn, she worked at Microsoft Bing. She holds a PhD in Statistics from Stanford University.
Facebook Recommendations
by Bradley Green, Jinyi Yao (Facebook)
Facebook Pages provide people the ability to connect to a wide audience. The goal of Page recommendations at Facebook is to connect people to the Pages that matter most to them, and to connect Pages with the most engaged audience possible. In this talk we will outline the history and development of recommendations atFacebook from social recommendations to content based recommendations.
Bradley Green is the engineering manager for Pages and Community recommender systems at Facebook. Since graduating from Georgia Tech he has worked on a diversity of relevance and ranking problems including search analytics at Microsoft, search and newsfeed ranking at Facebook. Jinyi Yao is the tech lead of pages recommendation at Facebook. Since graduating from Tsinghua University he has worked on a diversity of relevance and ranking problems for almost 10 years, including web document understanding, local search relevance at Microsoft, pages recommendation at Facebook.
Making Advertising Personal: Large Scale Product Recommendation at Criteo
by Pierre-Emmanuel Mazare (Criteo)
Behavioral retargeting consists in displaying online advertisements that are personalized according to each user’s browsing history. To this end, the selection of the products to display in the banner needs to be fast and accurate. In this presentation, we describe how Criteo’s stack mixes online and offline computations in order to run product recommendation in a few milliseconds at a large scale.
Pierre-Emmanuel Mazaré is engineering lead for the product recommendation systems at Criteo. Prior to joining Criteo in 2012, he was software engineer for Telenav, Inc., working on road traffic prediction systems. Pierre-Emmanuel holds a M.Sc. in applied mathematics and computer science from Ecole Polytechnique (France) and a M.Sc. in civil and environmental engineering from UC Berkeley (California).