Learning internal representations by error propagation
Parallel distributed processing: explorations in the microstructure of cognition, vol. 1
Proceedings of the fifth international conference on simulation of adaptive behavior on From animals to animats 5
Multiple paired forward and inverse models for motor control
Neural Networks - Special issue on neural control and robotics: biology and technology
Neural Networks - 2004 Special issue: New developments in self-organizing systems
| Hi-index | 0.00 |
This paper introduces the so-called "Forwarding Forward Model" network that explains how complex behavior can be learned and generated while its sensory-motor flow is hierarchically articulated. This model is characterized by a distributed representation of behavior primitives at each level, which contrasts with our prior models utilizing localist views. The model was examined through experiments using a 4-degrees of freedom arm robot with a vision system. The experimental results showed that behaviors can be generated both robustly and flexiblely going through the bottom-up and the top-down interactions between levels. The characteristics of the distributed representation are discussed. Our discussion is further extended to the phenomenological issue of subjective time perception. A novel idea for explaining the sense of" nowness" is derived by applying our idea of articulating experiences to Husserl's notions of retention and protention.