hotspot frontier focuses: deep learning of spiking neural network -- SNN Chinese
____It has been almost 150 years since human being discovered nerve cell for the first time in 1872. Though people have already had clear recognition to the bio-electric phenomenon of a single neuron, they still know little about the specific working mode of biological neural network. In today with such great scientific development, people still have no idea of how their nervous systems work specifically even for the simplest insect (e.g. an ant or a dragonfly) .
____It has been over twenty years since the first times boom on neural network at the end of 1980s. Traditional artificial neural network has been applied in many fields, however, with the deepening of research, its limitation and existing problems gradually appeared because the time coding has not been considered. Therefore, spiking neural network emerged at the right time, which has attracted more and more researchers attention.
____However, due to the complexity of spiking neutral network, people still have not found an effective learning algorithm till today and the research results presented here will change this situation. In other words, the results included an effective universal learning algorithm which is fit for multi-layer, multi-space, spiking neural network. Unlike traditional neural network learning algorithm (e.g. BP algorithm), this algorithm is more fit for the characteristic of biological neural network. Not only has it simulated the shaping process of neural synapse (connection), but it also has simulated the growth and disappearance of neural synapse (connection). My demonstration model has exhibited the training result of this learning algorithm, which has vividly and systematically demonstrated the activity of biological neural network.
____Note: There are mainly three reasons behind it: 1. Biologys neural system is usually composed of large quantities of neurons; 2. There exist extremely complex connections among neurons; 3. Neural activity is a dynamic and non-static process.
|Bionic Spiking Neural Network(Eng).rar||Introduction of relevant theories, models, and the demonstration programs.|
|Demo.rar||Intelligent system simulation demo program, in which, Demo 5 is worth paying attention.|
|Demo-A.rar||The deep learning simulation demo program for Appendix A.|
|Demo.cpp.rar||Source code of intelligent system simulation demo program.|
|Demo-A.cpp.rar||Source code of the deep learning simulation demo program for Appendix A.|
_Copyright ?font face="Times New Roman">2013 spikingnn.net All Right Reserved.
hotspot frontier focuses: deep learning of spiking neural network -- SNN