Recommendation Systems and Deep Learning at eBay

  • Broadway Room, Lerner Hall

RSVP to the Facebook Event!

Recommender systems have become an integral method for discovery of content on the web, such as music, movies, books, search queries, social media content, and consumer products. In the context of e-commerce websites like eBay, they can be a critical part of a user'sā€™ shopping experience, helping the buyers find the best products for them from among the millions that are available. A technique called collaborative filtering is the foundation of modern recommender systems, where behavioral signals such as item clicks and purchases are used to predict whether a user will find an item relevant. In this talk, we give an overview of standard collaborative filtering techniques, describe the challenges of applying collaborative filtering in a semi-structured marketplace such as eBay, and present how we are leveraging deep learning techniques to overcome the distinctive challenges of building recommender systems for eBay listings.

Speaker Bio
Daniel Galron is a research scientist & engineer working at eBay since 2014. He earned a PhD in computer science from NYU in 2012, where he worked on machine learning methods for machine translation.