Artificial Intelligence Conquers Poker fully

[Anonymous No.10805790]

- Dominated 6 player poker
- Ran on minimal amount of hardware (no GPU)

https://science.sciencemag.org/content/early/2019/07/10/science.aay2400

https://www.scientificamerican.com/article/ai-conquers-six-player-poker/

Multi-player Poker was conquered by AI.

Specifically, during live poker play, Pluribus ran on a machine with just two central CPUs and 128 gigabytes of memory. “It’s amazing this can be done at all, and second, that it can be done with no [graphics processing units] and no extreme hardware,” Sandholm says.

Two years later, he has proved himself wrong. Sandholm has co-created an AI program called Pluribus, which can consistently defeat human experts in six-player matches of no-limit Texas Hold’em poker. “I never would have imagined we would reach this in my lifetime,” says Sandholm, a professor of computer science at Carnegie Mellon University.

[Anonymous No.10805796]

>>10805790
What next, an AI that plays slots?

[Anonymous No.10805800]

>>10805796
poker was considered hard because of the psychology and necessity to bluff/lie/etc and vary strategy

[Anonymous No.10805804]

>>10805800
>psychology
It doesn't take that into account though. Just game theory based probabilities.

[Anonymous No.10805805]

>>10805790
>professor of computer science at Carnegie Mellon University.
>computer science

It goes in all fields

[Anonymous No.10805814]

>>10805790
>if he make this move
>search memory for the move that professional players told me to do
>win

Why CS cunts keep drooling over this autistic shit?

[Anonymous No.10805818]

>>10805814
weak associative reasoning, poor induction skills, the idea of causality is harder for people afflicted with this

[Anonymous No.10805824]

>>10805818
Ok, nerd

[Anonymous No.10806026]

>>10805814
>>search memory for the move that professional players told me to do
Okay, you have fundamentally not understood the point.
The amount of possible games is far too large to make that a viable strategy.

[Anonymous No.10806073]

Really glad deep meme networks were not a part of this. Hopefully we can start to move on from that shit.

>>10805814
https://science.sciencemag.org/content/early/2019/07/10/science.aay2400
>The core of Pluribus’s strategy was computed via self play, in
>which the AI plays against copies of itself, without any data
>of human or prior AI play used as input.

How does it feel to be a brainlet.

[Anonymous No.10806111]

>>10805790
This is bullshit, because poker is a game of skill AND luck, so either the program is really lucky, or the poker players are really bad.

[Anonymous No.10806130]

Now what I need is a poker machine that can play online for me and looks like a peeson. Loses just enough to be human but wins to get me some CA$H MONEY

[Anonymous No.10806143]

>>10806111
The human participants in the 5H+1AI experiment were Jimmy Chou, Seth Davies, Michael Gagliano, Anthony Gregg, Dong Kim, Jason Les, Linus Loeliger, Daniel McAulay, Greg Merson, Nicholas Petrangelo, Sean Ruane, Trevor Savage, and Jacob Toole. In this experiment, 10,000 hands of poker were played over 12 days. Each day, five volunteers from the pool of professionals were selected to participate based on availability. The participants were not told who else was participating in the experiment. Instead, each participant was assigned an alias that remained constant throughout the experiment.

[Anonymous No.10806604]

>>10805814
No you brainlet, it didn't brute force it.
The entire point of why all these AI engines have been working in recent years for the different domains they operate in has been because programmers have found ways to NOT have to brute force it.

[Anonymous No.10806613]

>>10806111
>This is bullshit, because poker is a game of skill AND luck, so either the program is really lucky, or the poker players are really bad.
Trips checked but I have no idea what point you're trying to make.
Luck can matter for a single hand. You don't play a single hand. e.g.:
>>10806143
>10,000 hands of poker were played over 12 days.
Over the course of 10,000 hands luck becomes irrelevant. That's the same concept behind why statistical approximations work. The more instances of the task are processed the less random noise plays a role in determining results.

[Anonymous No.10806835]

>>10805790
Soooooooooooooooooo....

does this mean online poker is literally only played by AI now or BANNED or can I still conquer online poker with a 100% winrate and earn infinite money?

[Anonymous No.10806857]

>>10805790
I feel like next step would be to create a single AI with superhuman capabilities in multiple different games and applications, while trying to reuse and integrate parts of the neural network as much as possible.

[Anonymous No.10807493]

>>10806857
deepmind kind of did this with 2 player boardgames

[Anonymous No.10808074]

>>10805790
If true, does that mean that their funding can be withheld now? After all, they can make any amount of money via poker...

[Anonymous No.10808225]

>>10808074
They can

Algorithm is being kept secret for this reason otherwise they would release. The authors wanna release it but can't due to the fact it would destroy online poker

[Anonymous No.10808235]

>>10808225
>nothingofvalueislost.png

[Anonymous No.10808381]

Online poker is played via statistics. I do many tables at once and most of it is automated. Playing safe and boring makes money. My algos would dominate human nature too and they're not AI.

Most online poker players do the same so it's really just my bot vs theirs but mainly splitting quarters from actual humans.

[Anonymous No.10808413]

>>10808074
>they can make any amount of money via poker...
t. has never played online poker
They've had rules against bots and aggressive measures for identifying and banning bots for a very long time now. Being able to beat any human players is a different skill from being able to make a program that can bypass whatever anti-bot measures the poker site has taken.

[Anonymous No.10808505]

>>10806130
Online poker already consist of bots at 99%. And remaining 1% of humans lose their money 100% of times.

[Anonymous No.10808516]

>>10808505
just bought in $100, messed around for a few hours. at $400+ now.

AI will never beat a human deep invested in truth. My bet range, my timing all changes depending on the variables of each unique situation. At the sametime i'm hyper aware of how others are reacting, keeping track of patterns and try to identify Male aggressive.


mediation + spirituality + truth

[Anonymous No.10808519]

>>10808074
Poker market is quite small compare to stock market for example. And stock market not so different from poker if you think about it.

[Anonymous No.10808698]

>>10806073
>which the AI plays against copies of itself, without any data
>of human or prior AI play used as input

impressive. can someone explain why he didn't just hook it up to online poker and make millions or maybe he already did and is only now releasing it.

[Anonymous No.10808704]

>>10808516

>AI will never beat a human deep invested in truth

So they said about chess, a couple decades back. See >>10806613 - poker and all other games effectively stop being random for these game computers.

[Anonymous No.10808804]

>>10805790
couldn't you just let this run on an online gaminbling site and rake in some cash

[Anonymous No.10808811]

>>10808225
I think it might have hinted on much better strategies than what is currently known.

[Anonymous No.10809690]

>>10808704
yes calculation wise, AI has me beat all day. A calculator has me beat also. I'm talking raw decision making in the moment.

[Anonymous No.10809697]

>>10809690
>raw decision making in the moment
There is no meaningful distinction between "decision" vs. "calculations" in this context.

[Anonymous No.10809716]

>>10806026
No it’s not.

[Anonymous No.10809720]

>>10806143
Wait, did they not play at a open table where they could see the other human players? Becuase that defeats the purpose

[Anonymous No.10809724]

>>10809690
ur literally retarded