While the global public still widely debate the ethics of using artificial intelligence, the AI sector continues to develop at a staggering pace. When talking about poker, AIs are already well ahead of their human counterparts and can beat even professionals at the world’s most popular card game. You’ll need a lot of poker tips if you want to try going against one.
These poker AI’s are referred to as poker bots and are serious business in the AI community. Their abilities are also advancing at an unprecedented rate due to the incredible skills of artificial intelligence to process information and learn from millions of sample poker games.
An AI is much faster at computing and disseminating information than humans ever could be. To put this into perspective, AI can cram 20,000 years of human work into a single week. This allows artificial intelligence to master games such as poker or chess much faster than humans.
Here is a look at the top three creators of AI poker bots in the world and some of their most successful bots.
University of Auckland Poker Bots
Based in New Zealand, the University of Auckland’s AI development team is currently developing highly advanced poker bots. Its most renowned creations at the moment include Casper and Sartre. Introduced in 2009, Sartre is a remarkably advanced poker bot that has competed in the yearly World Series of Poker Bots, garnering a lot of exposure.
Though not as advanced as some of the other programs on this list, the University of Auckland’s innovations are staples in the poker AI world and fascinating creations in their own right.
University of Alberta Poker Bots
The University of Alberta is famous for producing Loki, one of the most renowned poker programs in the world. It has been an industry leader for nearly two decades. Initially released in 1997, the program has advanced considerably over the past few decades. Under the leadership of Michael Bowling, the AI development team has consistently contributed to the development of this advanced poker AI.
Another of the University of Alberta’s bots, Polaris, was the first poker bot to beat human professionals in 2008. By 2015, the department also launched Cepheus, the first poker bot to solve the Limit Hold’em game.
The University’s latest bot, Deepstack, is capable of beating poker pros with ease. In 2017, it was considered the second-best poker player in the world and can use deep neural networks to emulate human intuition and learn on the go.
Additionally, Deepstack has significantly reduced game situation probability, allowing it to simplify certain scenarios much further than a human ever could. Deepstack’s average win rate is multiple times higher than what is considered good for a human professional.
Carnegie Mellon University Poker Bots
Last but certainly not least are the artificial intelligence programs made at Carnegie Mellon University, designed by a team led by Noam Brown and Tuomas Sandholm. The team has made various poker bots over the past decade, including Tartanian, Claudico, and the current flagship AI bots, Libratus and Pluribus.
Released in 2015, Claudico was tested against four of the world’s best human players in an extensive game at the Rivers Casino. In total, $170 million in virtual cash was bet over the course of 80,000 hands. Though the bot showed considerable promise, it was defeated by the combined efforts of the human players.
In total, Claudico was beaten only by a very slim margin of less than $733,000. Without a pooled strategy employed by the poker pros, it would have likely defeated the group.
In 2017, CMU launched its updated poker AI called Libratus. Libratus is responsible for delivering the most decisive blow in the human vs. AI poker war. Throughout 120,000 hands played against four of the world’s top professionals, Libratus didn’t disappoint.
Getting revenge for the defeat of his older sibling Claudico, Libratus crushed all four human players with ease. It won at an average rate of about 14.7 big blinds per 100 hands, beating the human team by a total of $1,766,250 in chips.
The results are even more shocking when the probability of humans actually playing better than Libratus was calculated at around 0.0001% - 0.54%. This, in theory, means that a human, even on their luckiest day, would still be no match for the world’s most powerful AI poker bot. Libratus cost the University $9.65 million to create and is 30,000 times faster than an average personal PC.
Additionally, the AI has over 274 terabytes of RAM and bases its own self-taught strategy on 15 million core hours of computation from the Pittsburgh Supercomputing Center. In an interview, Libratus’ co-creator Noam Brown stated that a perfected version of Libratus might be able to beat the original by an even wider gap than the poker bot already beat its human adversaries. Plans are underway to create just that – Libratus 2.0.
Although these are the three top development teams behind the new class of Poker AI, many more are also in the game. The AI field, in general, is seeing a rapid rate of growth. The cutting-edge technology may still be a few years from mass-market availability. However, existing AI has already defeated us at our own favorite pastime, and things are only looking to worsen for humans in the future.
What’s more important to note is that these poker AIs are just the tip of the iceberg for the artificial intelligence industry. Governments, sales platforms, and defense, finance, healthcare, and cyber-security sectors, are already looking at these AI poker bots as the heralds of future AI-driven tech. We just hope they all remember HAL and Skynet as a cautionary tale.
The whole point of using AI to enhance customer service is to make the customers feel recognized and respected.
Weak AI is going to have the opposite outcome every time.Read Full Article - https://bit.ly/2NPIv50
Mobile apps are continually upgraded and artificial intelligence is powering this evolution as more intuitive apps enter the market.
AI’s adjustable quality and learning capabilities allow businesses to satisfy the needs of customers while creating a personalized experience for every user.
Let’s look at how AI will change the world of mobile app development in this article.
If you're about to work on an Artificial Intelligence project and have still not decided about which programming language to use, here's a cue.
We've listed the six most common programming languages to facilitate Artificial Intelligence projects.
