innovator

Add idea


Calendar

«    July 2019    »
MonTueWedThuFriSatSun
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
 

 

Advert

 

Payment

 

Advert

 

Authorization

Стартап
DateDate: 12-07-2019, 07:47

“Over the past two decades, AI systems have mastered more and more complex forms of poker. However, all previous breakthroughs were limited to the participation of only two players, ”said the publication.
Compared to the paired games, games with a large number of players caused difficulties for the AI. To overcome their scientists have equipped the system with "abstract action" and "abstract information." These properties are needed to help the system reduce the number of options and highlight those needed to make a decision.
To do this, experts have developed a program called Pluribus, which has learned to play Texas Hold'em with its five copies. Playing time after time at random, Pluribus gradually determined which actions and with what probability gave the best results. Analyzing this data, the system gradually improved its strategy.
The self-learning method was previously used in zero-sum games for two, such as backgammon, Dota 2, Starcraft 2, and doubles poker. But as experts note, the algorithms used there differed significantly from Pluribus.
How did the experiment go
The computer game was evaluated in two formats: five players against Pluribus and one player against five copies of Pluribus.
Each of those who took part in the experiment earned more than $ 1 million at poker. The participants were players Jimmy Chow, Seth Davis, Michael Galliano, Anthony Gregg, Dong Kim, Jason Les, Linus Leliger, Daniel McAuley, Greg Merson, Nicholas Petrangelo, Shaun Ruan, Trevor Savage, Jacob Thule.
Every day, five professionals were selected for the game. However, they did not know the other participants. Each person was given a pseudonym, which remained unchanged throughout the experiment. It was possible to track the progress of the players. Between people, $ 50 thousand was divided to motivate them to play with full dedication. The players were guaranteed at least $ 0.40 per hand. The amount could increase to $ 1.60 per hand, depending on the outcome of the game.
In a match with five professional poker players or with five copies of Pluribus playing against one player, the AI for the 12 days of the experiment showed the best results in more than 10 thousand hands.





Add comment

Your name:*
Your E-Mail:*
Input code
from the picture*
Enable this picture to display the security code
reload, if the code cannot be seen