Back to LING 385
Lecture 9
Psychology and Math
- neural networks have many neurons
- a few of these neurons receive input from the environment, the others process incoming information
- some of the processing neurons give the output of the NN
- neurons in a layer take information from the earlier layer
- neural networks have "knowledge" in the form of two sets of numbers: Weights W and Biases b
- each processing neuron has a vector of Weights and one single bias.
- given inputs to the NN, W’s and b’s, we can calculate the output neurons
- this is called Forward Propagation and it performs recognition using knowledge!