Title :
Three-dimensional neural networks
Author :
Bricker, Jeremy ; Tzanakou, Evangelia Micheli -
Author_Institution :
Dept. of Biomed. Eng., Rutgers Univ., Piscataway, NJ, USA
Abstract :
The objective of this project is to construct and test a model of the brain, or of a specific region of it. As the brain is more complicated than a simple pattern-recognition perceptron, a dynamic, time-dependent, three dimensional artificial neural network (ANN) has been created for the purpose of most closely simulating the learning and recognition processes. It is of interest to research the effects of different stimulus patterns on learning, as well as different topologies of the neural connections themselves and damage to these connections and neurons. Considering the latter, lateral and feedback connections are implemented and tested along with the feedforward ones. The former, however, is yet to be investigated, but the facilities to explore it have been created
Keywords :
Hebbian learning; brain models; feedforward neural nets; neural net architecture; pattern recognition; unsupervised learning; ANN; brain model; damage; dynamic time-dependent three dimensional artificial neural network; feedback connections; feedforward connections; hebbian learning rule; lateral connections; learning; neural connections; neurons; pattern-recognition perceptron; recognition process; specific region; stimulus patterns; three-dimensional neural networks; unsupervised training; Artificial neural networks; Biological neural networks; Brain modeling; Joining processes; Network topology; Neural networks; Neurofeedback; Neurons; Pattern recognition; Testing;
Conference_Titel :
Bioengineering Conference, 1996., Proceedings of the 1996 IEEE Twenty-Second Annual Northeast
Conference_Location :
New Brunswick, NJ
Print_ISBN :
0-7803-3204-0
DOI :
10.1109/NEBC.1996.503248
