Séminaire Neuro : Thomas Trappenberg
Titre du séminaire et orateur
Probabilistic Machine Learning, Small Data, and Neurocognitive Robotics.
Pr. Thomas Trappenberg, Dalhousie Univ., Halifax (Canada).
Date et lieu
Vendredi 21 juin 2013, 14h30.
Université de Cergy-Pontoise, St-Martin 1, bât. A, 5ème étage.
Machine Learning made a considerable impact on technical solutions in recent years, including robotics. I will briefly review machine learning methods with an emphasis of relating different approaches in a probabilistic framework. I will then discuss two of my recent research interest.
The first is to understand complex learning systems that have to learn with limited data and bounded rationality.
The second area is that of using robotics to understand how the brain works. Within this later area I want to discuss dynamic neural fields and how they have been applied to cognitive robotics. I will also discuss their relation to Bayes filtering that is central for modern localization methods.