Title :
Areas vs. layers: biological aspects of neural network design
Author_Institution :
Inst. fur Neuroinf., Ruhr-Univ., Bochum, Germany
Abstract :
Based on the principles found in biological systems a neural network model is proposed that finds a robust and adaptive representation of signal space. The self-organizing sub-networks analyze different signal properties. Local dynamics are used for signal-driven adaptation of network parameters
Keywords :
image representation; neurophysiology; physiological models; self-organising feature maps; vision; adaptive representation; biological systems; local dynamics; neural network design; receptive field; self-organizing sub-networks; signal space; vision; Artificial neural networks; Biological neural networks; Biological system modeling; Biological systems; Chemicals; Neural networks; Neurons; Robustness; Signal design; Vector quantization;
Conference_Titel :
Neural Networks, 1995. Proceedings., IEEE International Conference on
Conference_Location :
Perth, WA
Print_ISBN :
0-7803-2768-3
DOI :
10.1109/ICNN.1995.488978