[Article.Ai] The “Richard Feynman” of artificial neural networks!

I call 3b1b the Richard Feynman of Artificial Neural Nets.

• His neural net series has got to be the clearest neural net series out there.
• I have been somewhat studying neural nets since 2015, and I have written a neural net from scratch, but watching his series made me realize I was simply writing these neural nets mostly from memorization or recall, rather than understanding. (sometimes my memory treats me well)

▼

▼

A nice tutorial:

I made a nice tutorial to wrap up his series here, inspired by, and requiring readers to first watch his neural net series, especially video 4.

• In my tutorial, I break down 3b1b’s neural net series, into three simple parts. (you can ignore the introduction without the math, and go straight to the portion with maths, where I advise people to watch 3b1b’s series before continue reading.)

▽

▽

▽

▽

Reference Summary:

• 3blue1brown’s artificial neural net series, with nice animations.
• My nice tutorial on quora.
• My tutorial in pdf format, with coloured formulaes instead of non-coloured ones.

▼

▼

Author:

I am a casual body builder, and software engineer.