machine learning

..

https://en.wikipedia.org/wiki/Machine_learning

This document is a sort of reference for me about machine learning topics.

2023-06-06

reference capture

• gradient descent is a way to train a neural network. Gradient descent requires that the model is expressed in a way that one can “take the gradient” of the model.
• Here’s a nice guide to machine learning from the perspective of a Clojure programmer: https://matthewdowney.github.io/clojure-neural-networks-from-scratch-mnist.html
• Neural networks from scratch with Numpy. I started following this guide, but never made it to the finish line. Good first impression.
• I think Jack Rusher wrote a tiny neural network-like thing in 10 lines of plain Clojure, but I can’t find it.
• Online gradient descent written in SQL was a thing I found in my Firefox history.
• Adrian Tofting recommends looking at:
  • Heterogeneous Graph Learning (pytorch-geometric)
  • Using Graph Neural Networks to Approximate Mechanical Response on 3D Lattice Structures (youtube)