Download here: Episode 8 - An Uncertain Revolution
In this episode we dive into the revolution brought on by the field of probabilistic robotics with Claudio Mattiussi who is Senior Researcher at the Laboratory of Intelligent Systems in Lausanne, Switzerland. We then launch a most "uncertain" competition to see how our listeners are able to cope with uncertainty in estimating the cleaning capabilities of our Roomba robot.
As a Senior Researcher at the Laboratory of Intelligent Systems at the EPFL in Lausanne Switzerland, Claudio Mattiussi has been looking into the world of evolutionary computation, neural networks and machine learning applied to tasks such as reverse engineering gene regulatory networks or designing electronic boards.Thanks to his experience with real-world applications and years in industry, Mattiussi has become aware of the need to deal with uncertainty, which is present in most environments and living beings. As a solution, he presents the probabilistic or Bayesian approach to perceiving the world, with a touch of history, philosophy and projection. Rather than being against good old fashion artificial intelligence (GOFAI), or Brooks' Behavior Based approach, he proposes the "uncertain" revolution using the probabilistic paradigm as being a compromise for the future.
Finally, he discusses how the probabilities can be used to make decisions on robot behavior using neural structures and evolutionary techniques.
For a detailed view on some of the subjects presented in this show, win the new book on "Bio-Inspired Artificial Intelligence: Theories, Methods, and Technologies" written by Dario Floreano and Claudio Mattiussi, out on the 30th of September 2008.
To make you apply your own probabilistic approaches to a concrete problem, we'll be asking you to guess (or compute) the percentage of dirt collected by a Roomba robot in its own "uncertain" environment. We're waiting for your vote by Wednesday, September 24th at 9AM GMT.
All the details for the competition can be found in this forum post.
Claudio Mattuissi's Homepage: http://lis.epfl.ch/~claudio/
The Laboratory of Intelligent Systems: http://lis.epfl.ch
Book on Bio-Inspired Artificial Intelligence: http://mitpress.mit.edu/catalog/item/default.asp?ttype=2&tid=11657