What is Simulated Neuron

Is a computer algorithm that simulates a neuron. The brain is made up of many neurons, an estimated 250 billion in the human mind. Each neuron is connected to a bunch of neurons going into it and it has a set of outputs leading to other neurons. These connections are called Axons. A neuron is normally either off or on and when a threshold is met the neuron will either fire a signal, turning on, or it will turn off. The key element is the threshold that must be met to activate. This is what is changes when we learn how to do something and is what we simulate in our machine learning algorithms via what is known as a sigmoid function. The function has a weight value associated with it that goes up and down until it gets to the correct level that matches the input training data. Normally we send data from our input into the neuron and see if it fires on the output (forward propagation). But, when we are in a training session, the information flows from the expected output backwards (back propagation) through the networks and the weights are adjusted.

Perceptions and New Realities for the 21st Century Learner

Jennifer (Jenny) L. Penland (Shepherd University, USA) and Kennard Laviers (Sul Ross State University, USA)