Hi all, I recently finished writing a series of articles on implementing NNs from scratch, trying to keep it as relatable as possible by demonstrating the maths (math, americans might say) in spreadsheets, pseudocode, and Rust. I really hope someone gets something out of this! Thanks :)
#1, a single neuron: https://www.splinter.com.au/2024/03/10/neural-networks-1/
#2, one neuron's gradient: https://www.splinter.com.au/2024/03/20/neural-networks-2/
#3, training: https://www.splinter.com.au/2024/04/22/neural-networks-3/
#4, multiple layers and backpropagation: https://www.splinter.com.au/2024/07/10/neural-networks-4/
Thanks for your time, I hope it's somewhat informative :)