Earlier quoted context omitted.
Strong players dont play "trap moves"
Who said anything about strong players?
> AlphaGo
Google reveals secret test of AI bot to beat top Go players
111–120 of 214 posts
Re: Google reveals secret test of AI bot to beat top Go players
#112Re: Google reveals secret test of AI bot to beat top Go players
#113Earlier quoted context omitted.
As computers are able to evaluate positions faster (and therefore deeper), the "godlike" tactics are dominating over human-style strategy. It used to be that computers played "computer-like" moves because they didn't understand the position. Now, they play computer-like moves because "understanding" the position isn't as important as just being able to see 25+ moves ahead. In a nutshell, positional play in chess is s…
> Now, they play computer-like moves because "understanding" the position isn't as important as just being able to see 25+ moves ahead. I don't think you can "see 25+ moves ahead" in Go. The branchout factor is just too big.
Well, stouset is talking about chess, but in Go, Monte Carlo Tree Search plays out all the way to the end of the game; it just does so in a much less exhaustive way, for exactly the reason you mention.
Re: Google reveals secret test of AI bot to beat top Go players
#114Earlier quoted context omitted.
I would also like to see how the strength is affected by computing resources. For example, one server, versus 10 servers, versus a room full of servers, versus an entire datacenter. I wonder how close this is to playing a perfect game. I would also like to know if two perfect players would always end up with a draw, or if they would each win 50% of the games.
Either the white player would win in 100% of matches, or the black player would win 100% of matches. Go rules prevent draws by giving Black a non-integer score bonus (this is called Komi, http://senseis.xmp.net/?Komi); By definition of "perfect play", a Black perfect player either always wins games, or a White perfect player always beats him.
Your reasoning is correct but
the bonus score (komi) is given to white, which is second to move and is behind in the race at building territory (the thing that gets scored).
Re: Google reveals secret test of AI bot to beat top Go players
#115Re: Google reveals secret test of AI bot to beat top Go players
#116Earlier quoted context omitted.
As computers are able to evaluate positions faster (and therefore deeper), the "godlike" tactics are dominating over human-style strategy. It used to be that computers played "computer-like" moves because they didn't understand the position. Now, they play computer-like moves because "understanding" the position isn't as important as just being able to see 25+ moves ahead. In a nutshell, positional play in chess is s…
> Now, they play computer-like moves because "understanding" the position isn't as important as just being able to see 25+ moves ahead. I don't think you can "see 25+ moves ahead" in Go. The branchout factor is just too big.
My comment was in a subthread about computers playing chess in a more human-like manner.
Re: Google reveals secret test of AI bot to beat top Go players
#117It would be good to start experimenting with handicap stones to understand how many stones stronger the bot is. http://senseis.xmp.net/?Handicap Also experimenting with a new ranking system beyond 9 dan.
I would also like to see how the strength is affected by computing resources. For example, one server, versus 10 servers, versus a room full of servers, versus an entire datacenter. I wonder how close this is to playing a perfect game. I would also like to know if two perfect players would always end up with a draw, or if they would each win 50% of the games.
It's not only about the hardware but also about the training data sets, the neural network topology...
Then, if something better than MCTS is found that will provide advantage too. Remember that those bots are hybrid solutions.
Re: Google reveals secret test of AI bot to beat top Go players
#118We know that Master can figure out what it would play. We also know that its predecessor had a model for what moves a human professional would be likely to play. What I would find truly fascinating is if Master could divide moves that it plays, which professionals wouldn't, into groups based on a similar internal categorization of the moves. And then see if human minds can look at any of groups and come up with a hum…
Your described method of understanding his moves might work for some moves, but there will inevitably be moves that are just a sum of so many different probabilities that are so far down the road that even looking at the result nobody would be able to actually recognize such a move on slightly different board. At some point, if this method would be used, it would probably require a second AI that would help understan…
Obviously "automatically extract the principles in an understandable format" is a long term and probably unachievable goal.
However there is hope that cluster analysis on data about the internal reasoning process can successfully identify groups of positions that "seem to share a common principle". Success in that is a first step towards lots of interesting things.
Re: Google reveals secret test of AI bot to beat top Go players
#119Earlier quoted context omitted.
When playing against a much stronger player, it's always very hard to figure out why they play tenuki (make a move somewhere else on the board, apparently uncorrelated to the current fight). When AI gets strong enough (and it seems like it has already), it will just tenuki everyone all the time, while winning. Sounds like exactly what's happening already. It's past the event horizon for human understanding.
"past event horizon for human understanding" I think you are too pessimistic and that "centaur Go" will evolve in the same way that "centaur chess" is, where a player (or team of players) with access to the software can outperform the stand-alone software. I guess my vote is for "intelligence augmentation" over AI, if only because opaque models/algorithms cannot benefit from human creativity.
Centaur chess players do not outperform the best chess engines. That is based on a misrepresentation of one of the first centaur chess tournaments. Human modified moves only reduce the strength of the best chess engines now. The benefit that was reported was one of tactics one centaur chess competitor had over other centaur chess competitors, not centaur chess vs stand alone chess engine.
It is a misunderstanding that has been breathlessly repeated by futurists and transhumanists for years.
Re: Google reveals secret test of AI bot to beat top Go players
#120It seems that AlphaGo has got even better since when it defeated Lee Sedol. This time, AlphaGo is undefeated against other top players. However, it also gives other players a chance to practice against AlphaGo and finds weaknesses in AlphaGo, which is super important for any competition.
It would be super-weird if AlphaGo became worse that the previous version.