Alexander Schwing

Assistant Professor
Department of Electrical and Computer Engineering
Coordinated Science Laboratory
University of Illinois at Urbana-Champaign

Office: CSL 103
eMail:

Deep Structured Models

Coming soon...

Distributed Structured Prediction (cutting plane, incl. latent variables)

A re-implementation of cutting plane learning for structured predictors suitable for distributed computing.
  • Latent Variable Version: (Feb. 20, 2014) Source Code
  • Standard Version: (Feb. 20, 2014) Source Code
  • CMakeLists file: File generously provided by Gellert Mattyus from DLR. For questions please directly contact

Distributed Structured Prediction (incl. latent variables)

Main Features:
  • latent variable models
  • parallelized for high-performance computing
  • sparse feature vectors
  • any form of higher order regions
  • arbitrary counting numbers
  • New release: (Feb. 09, 2014) Source Code v5.2
  • Previous release: (published upon request, not maintained/supported anymore) Source Code v3

Distributed convex Belief Propagation

Source code for our CVPR 2011 paper. Notes for fully general C++ implementation (higher order, arbitrary graph):
  • Dependency: MPI Implementation, e.g. OpenMPI, MPICH2, ...
  • Amazon EC2 tutorial.
  • A Matlab example is included.
  • Support for the UAI file format.
  • Tested on Windows (Win32,x64), Linux (x64) and Mac OS X (x64) operating systems as well as Amazon EC2 and the ETH Brutus cluster.
  • See included README for further details.
  • The downloadable distributed convex belief propagation algorithm has a destinct server task which loads the entire graphical model to its memory. This is intractable for really large models, where "large" is determined by the available server memory. We have other implementations waiting for your real challenges. Please don't hesitate to contact us.
  • convex BP File Format Description
  • UAI File Format Description
  • New release: (Sept. 18, 2011 - see CHANGELOG for details) Source Code v1.1

Random Forest

A non-adaptive C++ implementation of Random Forests used as a baseline in our CVPR 2011 paper.
  • Dependency: none
  • Operating system: Windows, Linux and Mac OS X
  • See README for further details
  • Source Code