Replicating The Brain
Positive and Negative Conditioning
Neural networks learn using a similar algorithm that the human brain utilizes named weighting algorithms. The human brain experiences three generalized situations; positive, neutral, or negative stimuli. When exposed to positive stimuli, neurons in the brain strengthen the associations between to the neurons involved in the exchange. Neutral stimuli yield minimal change, while negative stimuli rewire the neurons to avoid future occurrences of that particular stimulus.
Similarly, positive and negative conditioning can optimize neural networks in computing. During the initial learning phase of a neural network, if the output produced does not match the intended output, the network will create an error signal, readjusting the weights of nodes in the network of input nodes. In our hypothetical program, the neural network has just “learned” from its negative situation. Conversely, if a desired output matches an output from the network, weights and connections between nodes are improved. Positive and negative associations are both crucial factors in the success of both the human brain and neural networks.
Similarly, positive and negative conditioning can optimize neural networks in computing. During the initial learning phase of a neural network, if the output produced does not match the intended output, the network will create an error signal, readjusting the weights of nodes in the network of input nodes. In our hypothetical program, the neural network has just “learned” from its negative situation. Conversely, if a desired output matches an output from the network, weights and connections between nodes are improved. Positive and negative associations are both crucial factors in the success of both the human brain and neural networks.