Community Research and Development Information Service - CORDIS

FP7 MASH Logo

MASH

Project reference: 247022
Funded under: FP7-ICT

Massive Sets of Heuristics for Machine Learning [Print to PDF] [Print to RTF]

From 2010-01-01 to 2013-06-30

Project details

Total cost:

EUR 3 052 268

EU contribution:

EUR 2 308 999

Coordinated in:

Switzerland

Call for proposal:

FP7-ICT-2009-4

Funding scheme:

CP - Collaborative project (generic)

New tools for the collaborative development of very complex machine learning systems.

The MASH project aims at creating new tools for the collaborative development of very complex machine learning systems. Machine learning is concerned with the design of software able to learn from example. Since machine learning methods remain far from their biological counterpart in terms of performance, MASH will investigate a new strategy, by developing new theoretical tools and software to help large groups of individuals design large families of feature extractors. The idea is to combine several types of features developed by independent teams in order to improve performance. Because they exploit different sources of information, different modules mutually compensate their weaknesses.

Objective

This project aims at developing new machine learning methods relying on very large number of hand-designed heuristics, together with statistical tools to facilitate the design of these heuristics in an open and collaborative framework.We define an heuristic to be any algorithm processing raw inputs to produce values relevant to the problem at hand. This purposely very general definition encompasses techniques spanning from simple signal processing to symbolic modeling or locally trained predictors. Since we assume high performance can only be achieved by combining hundreds of such heuristics, we propose to develop them collaboratively, in a way similar to the successful development process of open-source software or collaborative encyclopedia.We will assess the performance of that strategy on the control of an avatar in a realistic 3D simulator and on the control of a real robotic arm, and we aim at creating a generic software platform usable on alternative applications.Hence, the key aspects of this proposal are to:- develop novel statistical techniques for prediction and goal-planning with a very large heterogeneous set of features,- develop statistical tools such as similarity measures in the space of features to help the design of very large sets of heuristics by many contributors,- assess the efficiency of this approach on a series of complex tasks in a realistic simulated 3D environment and with a real robot arm.The five partners of the consortium are from the fields of applied and theoretical statistical learning, reinforcement learning, artificial vision and robotics.

Related information

Documents and Publications

Open Access

Coordinator

FONDATION DE L'INSTITUT DE RECHERCHE IDIAP
Switzerland
RUE MARCONI 19
MARTIGNY, Switzerland
Administrative contact: François Fleuret
Tel.: +41277217739
Fax: +41277217712
E-mail

Participants

CESKE VYSOKE UCENI TECHNICKE V PRAZE
Czech Republic
ZIKOVA
PRAHA, Czech Republic
Administrative contact: Igor Mraz
Tel.: +420224352014
Fax: +420224357385
E-mail
UNIVERSITAET POTSDAM
Germany
AM NEUEN PALAIS
POTSDAM, Germany
Administrative contact: Regina Gerber
Tel.: +493319771080
E-mail
INSTITUT NATIONAL DE RECHERCHE EN INFORMATIQUE ET EN AUTOMATIQUE
France
Domaine de Voluceau, Rocquencourt
LE CHESNAY Cedex, France
Administrative contact: Mireille MOULIN
Tel.: +33 1 7292 5964
Fax: +33 1 7292 5936
E-mail
CENTRE NATIONAL DE LA RECHERCHE SCIENTIFIQUE
France
RUE MICHEL -ANGE
PARIS, France
Administrative contact: Gilles Pulvermuller
Tel.: +33 3 20 12 58 07
Fax: +33 3 20 63 00 43
E-mail
UNIVERSITE DE TECHNOLOGIE DE COMPIEGNE
France
Centre Benjamin Franklin, Rue Roger Couttolenc
COMPIEGNE, France
Administrative contact: N/A N/A
Tel.: +00 0 000000
E-mail
Record Number: 93563 / Last updated on: 2014-09-09