🎉 Celebrating 25 Years of GameDev.net! 🎉
Not many can claim 25 years on the Internet! Join us in celebrating this milestone. Learn more about our history, and thank you for being a part of our community!
Author
Hello Guys ,
I need to know , where can i find resources of programming AI in computer so that it can write hi nomi too me back and responds to me back in various conditions..
can somebody help with this , i find myself as an Advanced C/C++
Programmer.
Thanks..
Megahal I have been thinking on making a megahal dll to use with NPCs in a game, if anyone is interested, drop me a line, I already have started some of this, but right now I am working on pathfinding AI, and fixing some things in my ISO engine (structure mostly) and well, got to have a GUI before adding Hal.
3 trillions and one neuron :D
Or more...
My TicTacToe NN''s have over 3000 AN''s, and they don´t even understand how to block =/
I guess i should increase the number of layers or something, but that slows down the process
Or more...
My TicTacToe NN''s have over 3000 AN''s, and they don´t even understand how to block =/
I guess i should increase the number of layers or something, but that slows down the process
Sometimes less is more.
reduce the number of the layers and the count of input/putput Neurons. 9 for input, 9 for output and experiment some with one single middle layer.
The point is the more neurons you give for a simple problem, the more the network will specialize on certain situations and not learn the rules behind it. But you don''t want it to learn on how to react on a certain situation, you want it to learn the rule on how to react. And this can''t be archieved if you give too many neurons...
Hope this helps you.
reduce the number of the layers and the count of input/putput Neurons. 9 for input, 9 for output and experiment some with one single middle layer.
The point is the more neurons you give for a simple problem, the more the network will specialize on certain situations and not learn the rules behind it. But you don''t want it to learn on how to react on a certain situation, you want it to learn the rule on how to react. And this can''t be archieved if you give too many neurons...
Hope this helps you.
This topic is closed to new replies.
Advertisement
Popular Topics
Advertisement