Duke University Machine Learning

  • hartemink-jmlr-2012-image

    Hartemink group develops non-stationary dynamic Bayesian networks and models learning in songbirds

  • pca sfa adm gsk 2pcs color plot

    Engelhardt group infers population structure from genotypes

  • stratified2

    Mukherjee group stratifies spaces

  • pcbi 1002098 g004lrg

    Ohler group models Drosophila development

  • parr icra 2011 original

    Parr group uses a textured occupancy grid for simple robot localization

There is extensive, cross-departmental machine learning research going on at Duke University. Our faculty and students currently have projects that include graphical models, nonparametric Bayesian modeling, reinforcement learning, sparse approximation, compressive sensing, and manifold learning. Application areas include systems biology, fault detection, imaging, structural biology, genetics, cognitive science, and robotics. Find out more

Machine learning blog

Special Seminar: Jason D Lee, Exact Statistical Inference after Model Selection

February 25th 2014, 12:30-2:00, 330 Gross Hall Jason D Lee (Stanford) Title: Exact Statistical Inference after Model Selection Abstract: We develop a framework for post-selection inference. At the core of our framework is a result that characterizes the exact (non-asymptotic) distribution of linear combinations/contrasts of truncated normal random variables.…
Share |

First Machine Learning Seminar: Dave Blei (Princeton), Wed, Sept 18th

Dave Blei (Princeton) will give a talk entitled "Probabilistic Topic Models of Text and Users" in French 2231 on Wednesday, September 18th, 2013 at 3:30pm. There will be a reception afterwards in the iiD space on the third floor of Gross Hall. This is our kickoff talk in the…
Share |

Information Theory course by new faculty Galen Reeves

New faculty member Galen Reeves is teaching ECE 587: Information Theory, Fall 2013.Information theory is the science of processing, transmitting, storing, and using information. Pioneered by Claude Shannon in 1948 for problems in data compression and reliable communication, it is now relevant to a wide range of fields, including…
Share |

Duke Machine Learning Seminar Series

Beginning in fall 2013, Duke will be hosting a number of exceptional machine learning researchers on a bi-weekly basis for a seminar series. A tentative schedule for this fall includes Dave Blei, Sham Kakade, Tom Griffiths, and John Lafferty.
Share |

Peter Orbanz visiting Friday, February 15th

Professor Peter Orbanz from Columbia University is visiting Duke and giving the DSS Friday Seminar on Friday, February 15th. His work and interests are in representation problems and latent variable algorithms in Bayesian nonparametrics, and, more generally, in the mathematics behind machine learning.
Share |
  • mnist train2