Seminar: Google Seminar: Deploying Machine Learning @ Google: From Inspiration to Impact

SpeakerDoug Aberdeen
AffiliationGoogle
DateWednesday, 03 Oct 2012
Time14:00 - 15:00
LocationMPEB 1.03
Event seriesComputer Science Seminars
Description

"Why do we invest years of our professional lives in machine learning research? What difference does it make, to ourselves and to the world at large?" [Wagstaff, 2012]. Machine learning is making meaningful difference to the world every day, at least online. Google, Yahoo, Amazon, Microsoft, eHarmony and others use Machine Learning methods at massive scale to improve peoples' experiences on and off the internet. The aim of this talk is to give a practical insight into how Machine Learning is used at Google, from the development of algorithms through to deployment at scale. It will describe several disparate uses of Machine Learning in Google production, but the main case study will be the development of the Gmail Priority Inbox, one of the most user facing applications of Machine Learning at Google. It will also touch on user privacy and transparency issues that are special to Machine Learning. The talk will be light on theory and heavy on Machine Learning (and other) anecdotes.
Doug Aberdeen

Doug Aberdeen has been engineering Machine Learning at Google for the last five years, including spam detection, abuse detection, Gmail Priority Inbox, and most recently ad targeting. Before making the transition to industry he was a post-doc in the reinforcement learning and probabilistic planning communities. He's been a Geek since birth.

iCalendar csml_id_114.ics