TechBiiTechBii
  • Android
  • Computer Tips
  • How To Guides
  • SEO
  • WordPress
  • Content Writing
  • Tech News
Search
Categories
Reading: Artificial Intelligence in Poker: Have Computers Finally Caught Up?
Share
Font ResizerAa
TechBiiTechBii
Font ResizerAa
Search
Follow US
Tech Stuff

Artificial Intelligence in Poker: Have Computers Finally Caught Up?

Sidharth
Last updated: September 2, 2023 10:59 am
Sidharth
Published February 25, 2017
Share
6 Min Read

Artificial Intelligence (AI) has been one of the most controversial and most intriguing technology developments of the past several decades. The concept of AI is scary and fascinating at the same time, and it is hard to tell where the limits are.

Table of Contents
AI Provides a Lesson in HumilityLibratus Crushes Human Poker PlayersWhat Was Different about Liberatus?AI doesn’t need to “think” to solve pokerCan Math Substitute Intuition?

Until just recently, poker, or more specifically Texas Hold’em, was believed to be one such limit. Unlike other games successfully solved by computers, like chess or Go, poker involves a significant element of intuition and adaptability – something a computer, no matter how powerful, isn’t capable of.

AI Provides a Lesson in Humility
Libratus Crushes Human Poker Players
What Was Different about Liberatus?
AI doesn’t need to “think” to solve poker
Can Math Substitute Intuition?

AI Provides a Lesson in Humility

Can you really come up with a mathematical formula for “intuition?” Apparently, you can.

A team of scientists spent many hours working to develop a poker-playing Artificial Intelligence they dubbed Libratus. Libratus is, in fact, the younger brother of a similar project from a couple of years ago, named Claudico. Developed by the same team of scientists from Carnegie Mellon University, Claudico failed pretty hard in its match against top human poker players. But, this was only the beginning, and science wasn’t ready to throw in the towel just yet.

At the start of 2017, a new match was announced, featuring polished and improved poker-playing super-bot Libratus. Based on what we’ve seen in the past, it seemed like this would be another event that promises a lot but fails to deliver in the end. After all, how much better could Libratus be in comparison to its older brother?

Libratus Crushes Human Poker Players

Libratus claimed a decisive victory over its four human opponents, beating them across the span of 120,000 individual hands. The blinds were set at 50/100 for the entire 20-day match, which is a standard blind level at online poker sites like BetOnline. The AI ended up more than 1.7 million units ahead. In poker terms, it was winning 14.7 big blinds per 100 hands, which is a huge win rate, especially given the fact it was playing against a very tough competition.

What Was Different about Liberatus?

Going into the match, four human players (Jason Les, Dong Kim, Jimmy Chou, and Daniel McAulay expected it to be a relatively easy game. They were basing their belief on the 2015 Brains vs. AI match, where Claudico really didn’t represent a serious threat. It tried to adjust, but for the most part, it failed miserably.

Libratus was different, much different.

During the day, it would play the game, and during the night it analyzed the hands, looking for ways to improve. The four would quickly find out that finding ways to exploit the AI wasn’t going to work; they would find a weakness one day, but the next day, that weakness was gone. At the same time, the computer was adjusting to their playing styles, putting them in a position where they were the ones exploited and not the other way around.

AI doesn’t need to “think” to solve poker

The computer didn’t have actual intuition as such, but it didn’t really need it to win. Although poker may seem like a game of guts and feeling at first glance, it is, in fact, a game where the number of options is finite. It may be an enormous number, but it is finite. Once we accept this fact, what Libratus was able to do becomes much more understandable. It didn’t “read the souls” of its human opponents; it simply played billions of hands against itself, learning strategies for countless different situations.

Can Math Substitute Intuition?

The question we’re facing now is whether numbers and calculations can serve as a replacement for human intuition. Can such a complex, such a defining human skill, be enclosed in something so strict and so stern like math?

Apparently, it can.

What we call intuition is an extrapolation created from our past experiences, knowledge, and reasoning. If we are really rational about it, intuition is just a number of calculations packed up together, producing a certain result.

With an AI like Libratus, the ability to calculate outcomes of billions and trillions of different situations gives it an opportunity to develop a perfect “intuition” over time. With a finite number of possibilities and the best solutions for each, given enough play, Libratus could easily become an unbeatable poker machine.

So, while we may still be far from a machine capable of handling “matters of the heart,” with things like poker, where math and numbers are everything, we are definitely very close to having a perfect AI capable of beating any human player, no matter who they are or how good they happen to be.

Share This Article
Facebook Pinterest Whatsapp Whatsapp LinkedIn Reddit Telegram Threads Email Copy Link Print
Share
BySidharth
Follow:
Professional Blogger. Android dev. Audiophile.
Previous Article galaxy s nougat update [ROM] Samsung Galaxy S i9000 Lineage OS 14.1 Nougat Update
Next Article The Right Way To Plan And Track Email Marketing Campaigns
Leave a Comment

Leave a Reply Cancel reply

Your email address will not be published. Required fields are marked *

You Might Also Like

5 Quickest Tips to Play Poker without Making Use of Money

May 10, 2021
pari_match_app
Tech Stuff

What are Parimatch and the different features offered by Parimatch?

June 22, 2021
Tech Stuff

Top Ways To Sell Your Apple Products

May 26, 2021
Tech Stuff

7 Hotel Technology Trends To Adapt In 2023

November 30, 2022
FacebookLike
XFollow
PinterestPin
LinkedInFollow
  • Contact Us
  • Submit Guest Post
  • Advertisement Opportunities
Copyright © 2012-2024 TechBii. All Rights Reserved
adbanner
Welcome Back!

Sign in to your account

Username or Email Address
Password

Lost your password?