PinnedCategorical Cross-Entropy Loss and AccuracySo far, the model spits out a probability distribution so our loss function (can also be called Cost function) needs to reflect that, hence…May 4May 4
PinnedThe Forward Pass, or how the NFL Beat AI to the Punch by DecadesOK, not exactly, but the NFL did officially recognize the forward pass in 1906, and Frank Rosenblatt’s Perceptron didn’t come about until…Apr 2Apr 2
PinnedBig batches from little neurons grow.This text introduces the basic building blocks of neural networks by explaining how a neuron uses inputs, weights, and a bias to produce an…Mar 26Mar 26
A Little bit about Rectified Linear Activation Function (ReLU), and SoftmaxI’ll be discussing non-linear activation functions, because they’re sexy, at least compared to linear functions. I won’t be going overall…Apr 9Apr 9
Back(propagation) To The Future!During my time in the Navy I made liquid oxygen and liquid nitrogen (cool stuff-literally and figuratively). The ‘machine’ was a large gray…Jun 17, 2024Jun 17, 2024
(Yet another post about) Gradient DescentThe concept of gradient descent is easy to understand at a high level. Gradient Descent is like a hiking journey through a landscape of…May 22, 2024May 22, 2024
Differentiation and derivativesIt’s been said that no battle plan survives first contact, and writing is re-writing. With that said, my three-part series has morphed into…May 13, 2024May 13, 2024
My, what big features you have!I recently completed an AI/ML course through VetsinTech with Michael Galarnyk. He stressed that the most important feature in machine…Mar 24, 2024Mar 24, 2024