MASH logo
Login | Register

Massive Sets of Heuristics

The MASH project investigates the design of complex learning systems to increase the performance of artificial intelligence. To create such systems, we want to involve large groups of contributors from as many fields and backgrounds as possible, to program large and heterogeneous families of feature extraction modules.

Starting July 2013, MASH is now funded by the Hasler Foundation.

During its initial period 2010-2013, this project was funded by the Information and Communication Technologies division of the European Commission, Cognitive Systems and Robotics unit, under the 7th Research Framework Program.

The sub-project VELASH on object detection is funded by the Swiss National Science Foundation.

Read more >>

Latest news

The Factory contest prizes have been awarded

2013-12-02 13:57:47 by FrancoisFleuret

The Factory contest is over. The three winners are: Robert Lewis (UK), Elsa Fanchon (FR), and Frédéric Le Feurmou (FR).

Win one NEXUS 7 Android Tablet by participating in our contest!

2013-03-25 16:22:12 by Kanma

We are announcing the launch of a new contest in the fields of Machine Learning and Artificial Intelligence. The contest is aimed at all C++ programmers and does not require knowledge of either Machine Learning or Artificial Intelligence in general.

The contest begins today and will end on June 7th, 2013, after which three winners will each be awarded a NEXUS 7 Android Tablet (32Gb, 3G).

http://www.mash-project.eu/wiki/index.php/Factory_Contest

Contestants participate by submitting short C++ snippets, which we call "heuristics", designed to help a robot navigate a simulated 3D environment and solve a series of tasks (e.g. "Reach a red flag in a rectangular room"). Each of these heuristics receives as input the image corresponding to the robot's field of view, and should produce values ("features") which will then be used by a learning algorithm which attempts to mimic a "teacher".

The performance of a heuristic, or of a group of heuristics, is assessed by measuring how frequently the robot behavior coincides with that of the teacher.

Three prizes will be awarded:

  • Two for Best Accuracy. The contestants will be ranked for each task according to the accuracy they obtained, the winners shall be the two contestants with the highest average ranks.
  • One for Best Contribution. The submissions of the contestants ranked 3rd to 12th, according to their average rank, will be assessed by a jury composed of researchers involved in the MASH project. The jury will award the prize according to the generality and originality of the heuristics.

Each of the three winners will receive a NEXUS 7 Android Tablet with 32Gb RAM and 3G connectivity.

T-shirts and mugs with the MASH project logo will also be awarded to all contestants in the top-10 of average ranks.

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

The MASH Factory opens its doors!

2012-09-06 09:57:13 by Kanma

Write very simple code online and test it in less than 5 minutes! No registration, no download required.

Your task is to implement heuristics to help a virtual robot to navigate in a hostile environment and reach a flag.

You can test it by clicking on the Factory link on top of the page.

You can also look at the documentation, and listen to the dedicated screencast: