Matrices & Multiple Neurons
Scale single artificial neurons to parallel weight matrices, evaluating multiple criteria on the same input vector using matrix-vector multiplication.
Lessons in this Topic
Weight Matrix and Matrix-Vector Multiplication
Stacking multiple neuron weight vectors into a weight matrix to evaluate multiple criteria on the same input vector simultaneously.
Bias Vector and Multi-Neuron Linear Step
Independent baseline offsets for each decision criterion bundled into a bias vector to compute the multi-neuron linear step (z = Wx + b).
Multi-Neuron Forward Pass
Applying the element-wise Sigmoid function to the linear score vector to compute independent probabilities and parallel classifications.
Multi-Neuron Layer Architecture
Grouping parallel neurons into a unified layer, calculating multi-criteria forward passes, and examining the greenlight dilemma for network depth.
Matrices and Multiple Neurons In Practice
Manual problem sets covering weight matrix construction, matrix-vector multiplication, bias vectors, and multi-neuron layer forward pass classifications.