Asynchrony in a Distributed Modular Neural Network for Multimodal Integration

Y. Bouchut, H. Paugam-Moisy, and D. Puzenat (France)


Parallel Algorithms and Architectures, Neural Networks, Modelling and Simulation, Asynchrony


Starting from a cognitive point of view, a modular neural network has been previously designed for modelling a mul timodal associative memory simulating multisensory inte gration. This article presents a distributed implementation of the model, running on a network of computers. The pur pose of the parallel implementation is to exploit the asyn chrony of message passing to simulate temporal aspects in the model. Experiments in a virtual robotic environment highlight the interest of the model for a brain-like integra tion of multimodal informations, even in confusing context.

Important Links:

Go Back