This is a great way to show how Neural networks can learn, it starts from doing nothing (just standing) to completing one of the levels in a very fast-efficient time.
Time to learn to do this was around 24 hours - bear in mind this is a pretty crude design and not AI.
The "program" is based upon the following paper and topology: Evolving Neural Networks through
Augmenting Topologies
http://nn.cs.utexas.edu/downloads/papers/stanley.ec02.pdf
Please take time to watch it all as it shows how it behaves-learnt and the crux of what is involved with an Evolving Neural Network, initially starts showing speed it completes stage (up to 45s in video).
My only niggle is that he applies this to being comparable to natural evolution (in reality evolution is not as random and is more "smart" adaptation, viruses-pathogens and multi-stage parasites are easy examples on how the "classic" view of evolution is too simplistic but that is a discussion for other threads).
Anyway enjoy
https://www.youtube.com/watch?v=qv6UVOQ0F44
Cheers
Orb
Time to learn to do this was around 24 hours - bear in mind this is a pretty crude design and not AI.
The "program" is based upon the following paper and topology: Evolving Neural Networks through
Augmenting Topologies
http://nn.cs.utexas.edu/downloads/papers/stanley.ec02.pdf
Please take time to watch it all as it shows how it behaves-learnt and the crux of what is involved with an Evolving Neural Network, initially starts showing speed it completes stage (up to 45s in video).
My only niggle is that he applies this to being comparable to natural evolution (in reality evolution is not as random and is more "smart" adaptation, viruses-pathogens and multi-stage parasites are easy examples on how the "classic" view of evolution is too simplistic but that is a discussion for other threads).
Anyway enjoy
Cheers
Orb
Last edited: