Evolving creatures made with code (and their brains)

Channel:
Subscribers:
648
Published on ● Video Link: https://www.youtube.com/watch?v=oygWOUP8zoM



Duration: 9:27
591 views
20


Evolvio was originally created by carykh in Processing, which I was originally continuing on. It uses neural networks (specifically feed forward and NEAT learning) to make decisions for the creature. Cary Huang's version was just a simple feed-forward, however, mine is an LSTM (Long Short Term Memory) neural network AI which enables more complex decisions.

Live App: http://unknowngames.dx.am/project/evolvio/
Controls:
Try arrow keys. I, S, B. And probably other buttons I forgot about.
Oh well, they are in config.js if you're interested.

This is a web-based evolution simulator or game written in JavaScript language.

The source code is hosted on Github: https://github.com/evolvio4all/evolvio4JavaScript/
There are versions for other programming languages too: https://github.com/evolvio4all/

Music by Bensound — bensound.com







Tags:
evolution simulator
evolution
neural network
neural networks
large scale evolution simulator
evolv.io
simulator
carykh
evolvio carykh
neural network ai
air-n neural network
evolvio
walking evolution
walking creatures nerual network
evolution (software)
neural genetic programming
bipedal nerual network
neural net
network
evolvio lstm
neural
evolvio airn
neuroevolution of augmenting topologies
simulation
genetic algorithm
java
ai
Michael Denny
Timelapse