NEAT Algorithm Visually Explained

NeuroEvolution of Augmenting Topologies (NEAT) is a genetic algorithm (GA) for training artificial neural networks based on concepts taken from evolutionary biology. This animation was made in python using the package 'manim'. The voiceover was created using ElevenLabs.

Music by Vincent Rubinetti.

[nemongames]

It might be out of context, but you have the the voice of OmniMan

[floppaplatinum]

This is really great for beginners! Keep it up. I would love to see other games and configurations for the algorithm and how they could affect the simulation e.g. faster training time etc.

[kotreq]

Great explaination! Looking forward to more videos

[MrOnlineCoder]

Really great visuals, sad it has so small amount of views

[KX]

"more of a lucky coincidence than a well thought out strategy"
Or as it's called when Space-X boosters use this exact strategy intentionally in real life, a "suicide burn".

[revimfadli]

sounds quite unusual of NEAT to take this long for a simple problem

Great video though!

[xDLiker]

Do I understand that correct, that one always needs to define a reward function (in this case score) for this type of problem?

[fareahrahman]

Please make more videos on different algorithms.

[hazardousharmonies]

This is a superb demo David. Are you able to share your github code? Thanks for the video!