David Blei
2013 Regional Award Winner — Faculty
Current Position:
Professor of Computer Science and Statistics
Institution:
Columbia University
Discipline:
Computer Science
Recognized for: Contributions to the field of machine learning and Bayesian statistics.
Areas of Research Interest and Expertise: Statistical machine learning, large-scale text analysis, probabilistic modeling of massive data
David Blei
Biography:
PhD, Computer Science, University of California Berkeley
BSc, Computer Science and Mathematics, Brown University
David Blei is a pioneer of probabilistic topic models, a family of machine learning techniques for discovering the abstract “topics” that occur in a collection of documents. Over ten years ago, Blei and collaborators developed latent Dirichlet allocation (LDA), which is now the standard algorithm for topic models. Since then, Blei's research group has significantly expanded the scope of topic modeling. He and his students developed powerful methods for simultaneously analyzing documents and user behavior, such as scientists reading articles or lawmakers voting on bills. These new methods capture interpretable patterns of readership and can be used to recommend documents, characterize readers, and organize collections according to both content and consumption.
In addition to working on topic modeling, Blei has developed models of social networks, music and audio, images and computer vision, and neuroscience and brain activity. Recently, he and his students developed generic and efficient algorithms to fit a wide class of statistical models to massive data sets, changing the scale at which we can apply sophisticated methods for data science.
"My goal is to build new statistical tools for discovering and exploiting the hidden patterns that pervade modern real-world data sets, enabling experts to quickly summarize, navigate, and understand them."
Key Publications:
- Gopalan PK, Blei DM. Efficient discovery of overlapping communities in massive networks. Proc Natl Acad Sci U S A. 2013;110(36):14534-14539.
- Blei S. Probabilistic topic models. Communications of the ACM. 2012; 55(4):77–84.
- Hoffman M, Blei D, Paisley J & Wang C. Stochastic variational inference. Journal of Machine Learning Research. 2013; 14:1303–1347.
- Wang C & Blei D. Collaborative topic modeling for recommending scientific articles. Knowledge Discovery and Data Mining. 2011.
Other Honors:
ACM-Infosys Foundation Award in the Computing Sciences, 2013
NSF Presidential Early Career Award for Scientists and Engineers (PECASE), 2011
Alfred P. Sloan Fellowship, 2010
NSF CAREER Award, 2008
In the Media:
The Genius of David Blei. Fusion. Oct 27, 2014
David Blei | Embracing the Science of Teaching Topics to Machines. September 19, 2014
Avalanches of Words, Sifted and Sorted. New York Times. March 24, 2012
The Rise of Big Data. The New York Academy of Sciences Magazine. March 16, 2012.
ACM and Infosys Foundation Honor Leader in Machine Learning. Infosys. April 1, 2014