Massive sets of heuristics

The MASH project is a three-year research initiative to investigate the open and collaborative design of complex priors for Machine Learning.

This project is funded by the Information and Communication Technologies division of the European Commission, Cognitive Systems and Robotics unit, under the 7th Research Framework Programme. The sub-project VELASH on object detection is funded by the The Swiss National Science Fundation.

Research started in January 2010 and will be carried out until the end of 2012 in Switzerland (IDIAP), France (CNRS and INRIA), Germany (WIAS) and Czech Republic (CVUT).

Contact francois.fleuret@idiap.ch for more information.

Objectives

The goal of the MASH project is to create new tools for the collaborative development of large families of feature extractors. It aims at starting a new generation of learning software with great prior model complexity.

The project is structured around this web platform which will get fully operational in the coming weeks. It already comprises collaborative tools, such as a wiki-based documentation and a forum, and will soon integrate an experiment center to run and analyze experiments continuously.

The applications targeted by the project are classical vision problems, and goal-planning in a 3D video game and with a real robotic arm.

The scientific issues to be tackled along the course of the project are numerous, from standard machine learning questions such as learning and prediction with very large feature spaces and tight computational constraints, to original problems related to clustering in a functional space.

Consortium

The five institutions members of the project are:

• Idiap Research Institute in Martigny, Switzerland (coordinator).
• Heudiasyc laboratory of CNRS at University of Technology in Compiègne, France.
• Weierstrass Institute for Applied Analysis and Stochastics in Berlin, Germany.
• TAO Group of INRIA in Orsay, France.
• Center for Machine Perception of the Czech Technical University in Prague, Czech Republic.

People

• Abbet, Philip (developer, IDIAP)
• Beinrucker, Andre (PhD student, WIAS)
• Blanchard, Gilles (researcher, WIAS)
• Dubout, Charles (PhD student, IDIAP)
• Fleuret, François (researcher, IDIAP)
• Govaert, Gérard (researcher, CNRS)
• Grandvalet, Yves (researcher, CNRS)
• Hoock, Jean-Baptiste (PhD student, INRIA)
• Krämer, Nicole (researcher, WIAS)
• Lefakis, Leonidas (PhD student, IDIAP)
• Matas, Jiri (researcher, CVUT)
• Smutny, Vladimir (researcher, CVUT)
• Sokolovska, Nataliya (researcher, TELECOM ParisTech)
• Teytaud, Olivier (researcher, INRIA)