Heterogeneous Neural Networks for Adaptive Behavior in Dynamic Environments
Abstract
Research in artificial neural networks has genera1ly emphasized homogeneous architectures. In contrast, the nervous systems of natural animals exhibit great heterogeneity in both their elements and patterns of interconnection. This heterogeneity is crucial to the flexible generation of behavior which is essential for survival in a complex, dynamic environment. It may also provide powerful insights into the design of artificial neural networks. In this paper, we describe a heterogeneous neural network for controlling the wa1king of a simulated insect. This controller is inspired by the neuroethological It exhibits a and neurobiological literature on insect locomotion. variety of statically stable gaits at different speeds simply by varying the tonic activity of a single cell. It can also adapt to perturbations as a natural consequence of its design.
Cite
Text
Beer et al. "Heterogeneous Neural Networks for Adaptive Behavior in Dynamic Environments." Neural Information Processing Systems, 1988.Markdown
[Beer et al. "Heterogeneous Neural Networks for Adaptive Behavior in Dynamic Environments." Neural Information Processing Systems, 1988.](https://mlanthology.org/neurips/1988/beer1988neurips-heterogeneous/)BibTeX
@inproceedings{beer1988neurips-heterogeneous,
title = {{Heterogeneous Neural Networks for Adaptive Behavior in Dynamic Environments}},
author = {Beer, Randall D. and Chiel, Hillel J. and Sterling, Leon S.},
booktitle = {Neural Information Processing Systems},
year = {1988},
pages = {577-585},
url = {https://mlanthology.org/neurips/1988/beer1988neurips-heterogeneous/}
}