Principal investigator: Triesch
IM-CLeVeR aims to develop a new methodology for designing robot controllers that can: (1) cumulatively learn new efficient skills through autonomous development based on intrinsic motivations, and (2) reuse such skills for accomplishing multiple, complex, and externally-assigned tasks. During skill-acquisition, the robots will behave like children at play which acquire skills autonomously on the basis of “intrinsic motivations”. During skill-exploitation, the robots will exhibit fast learning capabilities and a high versatility in solving tasks defined by external users due to their capacity of flexibly re-using, composing and readapting previously acquired skills.
This overall goal will be pursued investigating three fundamental scientific and technological issues: (1) the mechanisms of abstraction of sensory information; (2) the mechanisms underlying intrinsic motivations, e.g. “curiosity drives” that learn to focus attention and learning capabilities on “zones of proximal development”; (3) hierarchical recursive architectures which permit cumulative learning. The study of these issues will also be fuelled by a reverse-engineering effort aiming at reproducing with bio-mimetic models the results of empirical experiments run with monkeys, children, and human adults. The controllers proposed will be validated with challenging demonstrators based on a single humanoid robotic platform (iCub). As a main outcome, the project will significantly advance the scientific and technological state of the art, both in terms of theory and implementations, in autonomous learning systems and robots. This overall goal will be achieved on the basis of the integrated work of a highly interdisciplinary Consortium involving leading international neuroscientists, psychologists, roboticists and machine-learning researchers.