Need advice for newbie
Posted: 27 February 2020, 06:36
Hi
I'm seeking for an advice as a non-programmer who is interested in learning programming. I have only very basic idea of what is coding. After I have heard about AlphaGo Zero, one thing pops up in my mind.
OK, now human's winning rate against AI is close to 0% in Go, but how about other games? If Alpha Zero learns how to play Lost Cities, is it going to win all the time? Unlike Go, most board games have some random aspect so I believe even Alpha Zero cannot win Lost cities all the time. I guess the winning rate might be 80% or something? What I want to study - perhaps life long dream - is to estimate the winning rate of many board games against AI. Go is 0% and the other extreme would be flip a coin game which is purely random so that AI's winning rate cannot be more than 50%. I believe AI's winning rate against human for most of the board games are in between 50% - 100% and it partially depends on how much random aspects are in each game.
To pursue my goal, I need to learn how to simulate board game, create decent AI and calculate it's winning rate against human. Is there anyone who can give me an advice where I should start? Is there any recommended programming language for this purpose? Can I use BGA studio for this purpose?
Thanks in advance.
I'm seeking for an advice as a non-programmer who is interested in learning programming. I have only very basic idea of what is coding. After I have heard about AlphaGo Zero, one thing pops up in my mind.
OK, now human's winning rate against AI is close to 0% in Go, but how about other games? If Alpha Zero learns how to play Lost Cities, is it going to win all the time? Unlike Go, most board games have some random aspect so I believe even Alpha Zero cannot win Lost cities all the time. I guess the winning rate might be 80% or something? What I want to study - perhaps life long dream - is to estimate the winning rate of many board games against AI. Go is 0% and the other extreme would be flip a coin game which is purely random so that AI's winning rate cannot be more than 50%. I believe AI's winning rate against human for most of the board games are in between 50% - 100% and it partially depends on how much random aspects are in each game.
To pursue my goal, I need to learn how to simulate board game, create decent AI and calculate it's winning rate against human. Is there anyone who can give me an advice where I should start? Is there any recommended programming language for this purpose? Can I use BGA studio for this purpose?
Thanks in advance.