Great video (as usual), very detailed and can be applied also to other types of NN. Enjoyed especially part about visualization of NN. Can't wait for part 2 with details about speciation. Thanks!
This is a nice explaination. In to intro, some of the shown networks seem to converge to a fitness of 3, mainly because it is giving a rather high value for the input (0,0). I'm cuurently facing the same issue in my implementation of NEAT. Do you have ideas to why this issue could occur? Thanks!
It's been an unsolved problem for many years now. With modern AI tech we might be able to find a stable solution to XOR, which is predicted to lead to many advances in science and industry.
How did you implement the lookup table? I was trying to recreate NEAT in python and I m a struglling to create it. How did you do it and can you think of a Python way do it as I understand you might have not used Python?
Great video (as usual), very detailed and can be applied also to other types of NN. Enjoyed especially part about visualization of NN. Can't wait for part 2 with details about speciation. Thanks!
Awesome, thank you!
only 1k views? this deserves waaaay more
looking forward for part 2!
working on it now.. it'll focus on speciation..
brilliant content, thank you :) so neat!
Glad you enjoyed it!
You can implement that with graph implementation?
This is a nice explaination. In to intro, some of the shown networks seem to converge to a fitness of 3, mainly because it is giving a rather high value for the input (0,0). I'm cuurently facing the same issue in my implementation of NEAT. Do you have ideas to why this issue could occur? Thanks!
What is the application of solving XOR?
It's been an unsolved problem for many years now. With modern AI tech we might be able to find a stable solution to XOR, which is predicted to lead to many advances in science and industry.
@@fappylp2574 Just wait until we solve the half-adder ...
How did you implement the lookup table? I was trying to recreate NEAT in python and I m a struglling to create it. How did you do it and can you think of a Python way do it as I understand you might have not used Python?
Maybe the disabled ones could be faint semi-transparent, so you can see them if you try hard enough but they're not distracting most of the time?
Why no code?
Uumpf bit of an overkill don't ya think?
is your code available on github or anything ? thx
Awesome o/
I did all this, but evaluating mnist takes too damn long.