quote:

In October 2015, a computer Go program called AlphaGo, powered by DeepMind, beat the European Go champion Fan Hui, a 2 dan (out of 9 dan possible) professional, five to zero.[23] This is the first time an artificial intelligence (AI) defeated a professional player.[5] Previously, computers were only known to have played Go at "amateur" level.[23][24] Go is considered much more difficult for computers to win compared to other games like chess, due to the much larger number of possibilities, making it prohibitively difficult for traditional AI methods such as brute-force.[23][24] The announcement of the news was delayed until 27 January 2016 to coincide with the publication of a paper in the journal Nature describing the algorithms used.[23]

quote:

AlphaGo is scheduled to challenge South Korean professional Go player Lee Se-dol, who is ranked 9 dan,[5] with five games taking place at the Four Seasons Hotel in Seoul, South Korea on 9, 10, 12, 13, and 15 March 2016,[13][14] which will be video streamed live.[15] Aja Huang, a DeepMind team member and amateur 6-dan Go player, will place stones on the Go board for AlphaGo, which will be running through Google's cloud computing with its server located in the United States.[16] The match will adopt the Chinese rules with a 7.5-point komi, and each side will have two hours of thinking time plus three times of 60-second byoyomi.[17]

The winner will get a $1M prize. If AlphaGo wins, the prize will be donated to charities, including UNICEF.[18] Besides the $1M prize, Lee Se-dol will receive at least $150,000 for participating in all the five games and additional $20,000 for each win.[17]

echinopsis posted:

go? like magic the gathering?

1488 posted:

the robit is just going to cheat, its the only way they know to beat humans

graph posted:

no as in othello but its different and idk how

Symbolic Butt posted:

9x9 go is super boring, it's the exact same rules as regular 19x19 go but it's essentially a tactical game much like say hex

some weird poo poo happens when the board gets bigger, tactics are still important but it also gives rise this kind of vague strategical notion of ~~influence~~ which is probably what tef probably meant by "mystical bullshit".

imo it's a super interesting thing that it's hard to learn without some weird human intuition, this is pretty much what was considered a huge to challenge to AI in the early 2000s: make the computer grasp the notion of influence much like a human would

better computers and monte carlo algorithms made the jump from a computer playing like a complete retard to a semi-competent player. the dream of AI thinking like a human didn't realize with go like people were predicting a decade ago but it's still a cool thing I guess yospos birch

Symbolic Butt posted:

addendum: you can be super great at calculating moves in a local scale (like a computar) but it's impossible to crunch every variation on a global (whole board) scale so humans rely on this influence intuition

Symbolic Butt posted:

chess also has some aspects where a human player still have clear better judgment than a computer engine (like judging board positions) but it's less effective and flashy than what happened with go

and also pretty much reserved to dudes like magnus carlsen

Symbolic Butt posted:

I sperged about this board game in the terrible programmers thread so I'll quote those here

graph posted:

no as in othello but its different and idk how

Space-Pope posted:

this came up on the stream

echinopsis posted:

why can't the computer crunch? is there literally that many iterations that modern huge millioncore computers still can't brute force it?

echinopsis posted:

why can't the computer crunch? is there literally that many iterations that modern huge millioncore computers still can't brute force it?

Testiclops posted:

there are apparently
208168199381979984699478633344862770286522453884530548425
639456820927419612738015378525648451698519643907259916015
628128546089888314427129715319317557736620397247064840935
legal positions, which are a lot, op

echinopsis posted:

why can't the computer crunch? is there literally that many iterations that modern huge millioncore computers still can't brute force it?

Symbolic Butt posted:

a good rule of thumb that I learned with Peter Norvig (my computerman hero) with dealing with brute force nowadays is:

brute forcing 1 million to 100 million states (10^6 - 10^8) is trivial, modern computers can do it in a matter of seconds, maybe a minute

at billion states (10^9 - 10^11) it starts getting a little harder, like maybe it'll be like 20 minutes to 3 hours, something like that. if you really want an answer in the worst case you could leave your computer running it for a week? no big deal

now trillion states (10^12 - 10^14)... you're just hosed. you'd need to leave your computer running it for hundreds of years, it's just hopeless. even with the best computers on earth at this point it's still tough as poo poo with a pure brute-force approach.




this board game has around 10^170 possible states

Space-Pope posted:

a mobile app when you want to play (or watch) on the toilet

echinopsis posted:

Symbolic Butt posted:

a good rule of thumb that I learned with Peter Norvig (my computerman hero) with dealing with brute force nowadays is:

brute forcing 1 million to 100 million states (10^6 - 10^8) is trivial, modern computers can do it in a matter of seconds, maybe a minute

at billion states (10^9 - 10^11) it starts getting a little harder, like maybe it'll be like 20 minutes to 3 hours, something like that. if you really want an answer in the worst case you could leave your computer running it for a week? no big deal

now trillion states (10^12 - 10^14)... you're just hosed. you'd need to leave your computer running it for hundreds of years, it's just hopeless. even with the best computers on earth at this point it's still tough as poo poo with a pure brute-force approach.