The MASH project
The MASH project: An open platform for the collective development of future extractors by Francois Fleuret - Machine Learning Summer School at Purdue, 2011.
The MASH project is a European research initiative that aims at developing novel tools for the design of complex machine learning systems.
The project's main objective is to create an open-source web platform allowing the collaborative development of complex families of image feature extractors for scene understanding and vision-based goal-planning algorithms. Performance will be measured on standard image classification and object detection data sets, and using a real robotic arm, and a 3d simulator for goal-planning tasks. I will present the motivations behind the project, the scientific and technical challenges to address, and the current status of the platform. I will then show how to participate and some early results.
|The MASH project|
Machine Learning Summer School at Purdue, 2011