Learning and communication via imitation: an autonomous robot perspective

Abstract

This paper proposes a neural network architecture designed to exhibit learning and communication capabilities via imitation. Our architecture allows a "proto imitation" behavior using the "perception ambiguity" inherent to real environments. In the perspective of turn-taking and gestural communication between two agents, new experiments on movement synchronization in an interaction game are presented. Synchronization is obtained as a global attractor depending on the coupling between agents' dynamic. We also discuss the non-supervised context of the imitation process and we present new experiments in which the same architecture is able to learn perception-action associations without any explicit reinforcement. The learning is based on the ability to detect novelty or irregularities in the communication rhythm.

Emotion understanding: robots as tools and models

“…We were able to show that this kind of mechanism can be used to teach an arbitrary set of sensory motor rules to a computer without giving it any reward (Andry et al 2001). It was the frequency of the trials of the human demonstrator that was directly used by the computer to infer if it has or not to reinforce its current behavior.…”

“…A mechanism rooted in an internal "value system" would be needed to provide internal signals regarding the "positive" or "negative" qualities of actions and stimuli, giving them a meaning with respect to the values, needs and goals of the robot beyond a metaphoric use of the terms "pain" and "pleasure" to refer to negative and positive reward, and allowing to learn appropriate valenced reactions to them. See (Andry et al 2001), (Cos Aguilera et al 2003), and Section 4.2 of this chapter for initial solutions in this direction.…”

Toward learning autonomous pallets by using fuzzy rules, applied in a Conwip system

“…They consider the learning of global data for faster convergence and using agents for local data to achieve higher accuracy. All in all, as Dorigo and Colombetti [35] and Andry et al [36] referred to, learning is a mean of autonomy achievement.…”

Autonomous Development of Social Referencing Skills

“…After visuo-motor learning (learning between the extremity of the arm and the proprioception), several positions in the workspace are reached by the robot arm [1]. One visual position corresponds to one or several motor configurations (e.g attractors).…”

“…If a rhythmic interaction between baby and mother involves positive feelings and smiles (positive reward), a social interaction interuption involves negative feelings (negative reward). In our case (following [1]), the rhythm is used as a reward signal. It provides an interesting reinforcement signal to learn to recognize an interacting partner(face/non face).…”