Developmental Neuroscience and Neural Networks Development of neuroscience

Developmental Neuroscience and Neural Networks • Development of neuroscience ∙ Brain research is to clarify human's identity ∙ Be formed Neural system by connected several neurons ∙ Nervous system pass stimulation of the external environment to the brain ∙ Deliver signals to peripheral organs ∙ Using Microelectrode technology, To record single neuron's activity ∙ Complex behavior is participated by several sites of brain 1

• Computers and the Brain ∙ Neural network is a collection of neurons connected to various forms ∙ Neural network simulate what happens in nervous system. ◦ Difference between Computer and Brain Computer Brain Calculation speed 10 million / s Hundreds / s Work to how Serial Parallel Output Exact Variable Design to how Sophistication Biological evolution Configuration Hardware and Software Not a clear distinction Result Optimization calculations Optimizing recognition and understanding 2

• Development of Neural Networks ∙ Designed to operate in parallel computers ∙ Using probabilistic fuzzy logic AI program development ∙ Neural network express inputs and outputs set by the designers ∙ Neural networks are resistant to damage ◦ Parallel Distributed Processing Neural Network - Connection of simplified neurons Perceptron - Simple neural network consist by monolayer Connected to the multilayer, solve exclusive-OR logic Repeated over a finite number of times ⇒ Learning the correct actions 3

• Interactive activator model ∙ Using backpropagation algorithm (Neural System Identification) ∙ Composed of Input layer, hidden layer and output layer ∙ Connections exist only prospective way • Net talk ∙ Describe various aspects of learning to reduce ∙ Important competitive learning ⇒ Suppress output of other devices by the largest output device 4