The Department of Statistical Science at SMU is pleased to announce that Dr. Robert M. Bell will be our second Bill Schucany Visiting Scholar.
Robert Bell received his Ph.D. in Statistics from Stanford University. He spent about twenty years at RAND doing public policy analysis before moving on to the Statistics Research Department at AT&T Labs-Research. A few years ago he left Bell Labs for Google, from which he retired last spring. In 2009, he was a member of the team that won the $1 million Netflix Prize for building the most accurate algorithm for predicting Netflix customer movie ratings. He has served on the Fellows Committee of the American Statistical Association, the board of the National Institute of Statistical Sciences, the Committee on National Statistics, and several National Research Council advisory committees studying statistical issues from conduct of the decennial census to airline safety. His current research interests include machine learning methods, analysis of data from complex samples, and record linkage methods.
Dr. Bell will give two public lectures while visiting SMU on October 19 – 20, 2017:
Lessons from the $1,000,000 Netflix Prize
October 19, 2017, 7:00 pm, 133 Fondren Science Building, SMU Campus
In October 2006, the DVD rental company Netflix released more than 100 million user ratings of movies for a competition to predict new ratings based on prior ratings. The size of the data (over 17,000 movies and 480,000 users) and the nature of human-movie interactions produced many modeling challenges. One allure to data analysts around the world was a $1,000,000 prize for a team achieving a ten percent reduction in root mean squared prediction error relative to Netflix’s existing algorithm. Besides producing a photo finish worthy of a movie, the 33-month competition spurred numerous advances in the science of recommender systems and machine learning, more generally. After describing some of the techniques used by the leaders, I will offer lessons and raise some questions about building massive prediction models. This is joint work with former colleagues at AT&T Labs-Research, Chris Volinsky and Yehuda Koren.
Diverse Applications of Probabilistic Record Linkage
October 20, 2017, 11:00 am, 190 Crow Building, SMU Campus
The classical usage of probabilistic record linkage is to link common entities between two files using multiple, imperfect linkage variables like name, address, and birth date. The technique handles linking variables that are non-unique and/or error prone. The methodology also applies to many unexpected problems, invariably with interesting twists. I will illustrate with examples from some or all of public health, the decennial census, survey data collection, and criminal justice.