Principles for the co-development of sensorimotor internal models

Thumbnail

Event details

Date 09.12.2013
Hour 15:0017:00
Speaker Prof. Alexandre Bernardino and Prof. José Santos-Victor
Bio: Alexandre Bernardino and José Santos-Victor are professors at the Institute for Robotics and Systems, of the Instituo Superio Técnico in Lisbon, Portugal.

Both got their M.Sc and PhD from the IST in Lisbon.

Alexandre Bernardino's research interests focus on the application of computer vision, cognitive science, control theory and machine.

José Santos-Victor's research interests lie in the areas of Computer Vision and Vision-based Navigation of Robotic Systems.
Location
Category Conferences - Seminars
In this talk I will overview work done in the past few years at Vislab on learning internal predictive models. I will briefly introduce the concepts of expected perception and objects' affordances and stress the importance of internal models in predicting the consequences of the robot actions in its perception of the environment. Then I will present a methodology to simultaneously learn sensorimotor maps and the underlying sensor and motor signal representations, illustrating that general principles of prediction error minimisation and simplicity can lead to sensorimotor structures with similarities to biological systems.

Practical information

  • General public
  • Free

Organizer

Contact

Event broadcasted in

Share