Developmental Robotics
Generating plausible models for the processes underlying children’s development in the first years of their life is a challenging scientific issue at the crossroads of neuroscience, learning theories and developmental psychology. Children seem to acquire new know-how in a continuous and open-ended manner. A large amount of work describes how new skills seem to build one upon another, suggesting a continuum between sensory-motor development and higher cognitive functions. But very few plausible low-level mechanisms exist to explain how such skills emerge or self-organize.
Studying development is intrinsically difficult because of the complex interplay between embodiment, learning mechanisms and environmental dynamics. A relevant integrative approach can be pursued by viewing development as a complex system the dynamics of which can be studied with embodied models. In order to capture part of the open-ended nature that characterizes children?s development, we design new biologically-inspired architectures to control autonomous robots. In particular, we conduct research on motivational principles that can drive a robot to continuously try to master new know-how. The aim is to construct engines implementing such general capacities as “curiosity”, thus producing generic attention mechanisms with a minimum of preprogrammed biases.
This approach might not only help us understand the mechanisms underlying human development, but it might also provide radically new techniques for building intelligent robots. Indeed, as opposed to the work in classical artificial intelligence in which engineers impose pre-defined anthropocentric tasks to robots, the techniques we develop endow the robots with the capacity of deciding by themselves which are the activities that are maximally fitted to their current capabilities. Our developmental robots autonomously and actively choose their learning situations, thus beginning by simple ones and progressively increasing their complexity. No tasks are pre-specified to the robots, which are only provided with an internal abstract reward function. For example, in the case of the architecture which we developped, this internal reward function pushes the robot to search for situations where its learning progress is maximal.
Keywords : developmental robotics, epigenetic robotics, intrinsic motivation, curiosity, values, development, intrinsically motivated reinforcement learning, autonomy, epigenetic robotics, behavior, developmental trajectory, complexity, active learning.
Members: Frédéric Kaplan, Pierre-Yves Oudeyer, Verena Hafner
