---- Q: Briefly describe the perceptron learning algorithm that we described in class. ---- A: We discussed the 'back propagation learning algorithm'. This learning algorithm takes a neural network with known layers and known activation function, but unknown weights, and it takes sample inputs and outputs. Its goal is to use the sample inputs and outputs to learn what the unknown weights should be. It does this by first guessing values for the weights. Then it runs the inputs through the neural network, from which it obtains some outputs. These outputs are then compared with the given sample outputs, and the error between them is calculated with the loss function. The algorithm uses these calculations to adjust the values of the weights. This process repeats until satisfactory weights have been learned. ---- Answered by: Daniel Trebe, Alex Gilham, one other guy, and Cuauhtemoc Sanchez.

-- Practice Final #10

Post your names, guys, so I can add them to the OP.

-- Practice Final #10

I, Daniel Trebe, and Alex Gilham also worked on this and problem 9 alongside one other guy whose name I unfortunately do not know