May 1, 2026

Built single layer and single perceptron machine learning models

Implemented linear regression and classification problems from scratch using numpy and visualizing using matplotlib

Implemented a multilayer neural network to solve a classification problem

Learned how to use calculus for back propagation in training a neural network with simple layers(input and output, no hidden layer) and multiple perceptron for both linear regression and classification problems

Learned how to use calculus for back propagation in training a neural network with multilayer(a single hidden layer)

Learned about Newton's method and its advantages and disadvantages and the comparison between this and the gradient descent method

Learned about Hessian and Concavity

Took a sneak peak into how beautiful neural networks actually are and how they can't exist without linear algebra and how they can't learn without calculus. Anticipating what the role of probability and statistics will be on machine learning and neural networks(already seeing how it would play out but yeah, would be great to learn it in the next course)

N/A

The course basically did this for 2 layer network with maximum of two perceptrons per layer, my Mathematical mind is already thinking of the generalization of this. I would need some time to study the general case, would be monstrous but yeah...

Would probably start the Mathematics for ML(Probability and Statistics) but I have a couple other things to do tomorrow, I need to work on my Zama project, clean up some past CKB projects and maybe build a new thing on CKB and I also have to work on my project for Kite AI hackathon

Might need to rethink the whole Saturday thing, I only wanted to be taking a full day break on Sundays but some other things I have to do can be done on Saturday so I won't have to do any work on Sundays.

Completed course Mathematics for ML(Calculus)

Calculus Notes: Calculus.pdf