Key Research Activities

KRA 1:
Representation of Things and Actions
Define invariant, multi-sensory representations of objects through actions performed on them and processes by which these representations can be derived, refined and learned.

How OACs arise?
Derive sensory representations of actions and combine actions and objects into meaningful entities to create first early object-action complexes.

How OACs evolve
How more complex OACs can arise from the early OACs and how during this process Novelty and, hence, Creativity occurs.

Language, Communication and Interaction
Develop interaction and communication from shared actions and map OACs onto a grounded symbolic representation.

Key Research Activities

Memorization, Learning and Underlying Theoretical Concepts
Develop different learning and inference techniques and derive measures for predictability, complexity and contingency in the context of a closed loop perception action system.

Decision making, planning and evaluation (CSIC)
How to drive an agent to both minimize contingencies and to maximize rewards in the presence of uncertainty. Develop decision making and planning algorithms in the context of OACs for continuous perception and action spaces.

Cognitive Robotic Systems: Integration and Control Architecture
Design and develop a robotics system and integrate the theoretical and algorithmic results from other KRAs into a complete technical cognitive system.

IST-FP6-IP-027657  © PACOPLUS 2006-2010. All rights reserved.   E-mail: Webmaster
PACO-PLUS is funded by the European Commission through the Cognition unit, Information Society and Media DG.