TitanbetPoker – Now that there is a computer that can beat humans at Texas Holdem poker, should players fear fraud in casinos or is this mostly a fantasy which has become exaggerated and warped into a kind of urban legend? Steve Ruddock says that there is no reason to fear the computer Cepheus which has in essence already solved the poker game, but is he really right about this? Let’s have a closer look at the field of artificial intelligence, AI, to understand a bit more about it and what we might expect of future computers and robots…
Artificial Intelligence
The term artificial intelligence was coined back in 1956 by researcher John McCarthy during the Dartmouth Conference. It was during this conference that the main mission of AI was outlined. Any computer program that could do something that we would normally considered exclusively human would be considered an AI program. The focus was on the completed result and not so much on the means, meaning the structures and processes leading to successful action do not have to look like the thought processes of a human being. This is still an important point to keep in mind. When speaking about artificial intelligence one could come to the conclusion that it has to function in the very same way that a human being’s intelligence does. However, this has never been the case in the history of the development of AI.
When talking about the field of AI it is important to define just what is meant by intelligence. The definition, according to Stanford University, of the type of intelligence that can be applied to computers is that it is the “computational part of the ability to achieve goals in the world.” Now, these goals can be very diverse. With the poker Cepheus computer the goal is obviously to beat real players by making the correct decisions based on many different factors such as the cards that are dealt. Thus, AI can be summarized as the science of making computer programs that are intelligent. However, it is very important to understand this doesn’t mean that the computers will ever achieve human intelligence. The programming looks in depth at how we humans solve problems and then creates intelligent programs that try to mimic this.
Fears of AI
Which #ArtificialIntelligence Categories Are Seeing the Most Innovation? http://t.co/0UBbS3yb6r http://pic.twitter.com/RisGguPyg0
— Khalid Hamdan (@KhalidHamdan0) June 14, 2015
There is a lot of talk about AI and also a lot of fear surrounding this subject. The discussion may bring up half-joking fantasies of computers killing humans or at least taking away our jobs. The latter idea is easy for many people to relate to since we’ve already seen how developing technologies are making many previous job assignments unnecessary. So the fears around the smart computers might have some grounds to them. However, the question is whether we are being influenced somewhat by movies like Terminator and other stories from Hollywood where the robots come to life and take over our world. Looking at the current state of AI in different fields this is far from reality. And even if there arerobots that look like humans and that can handle tasks such as the check out in stores these robots are not really like humans and it is hard to believe that they ever will be. So while we might see pretty robots with hundreds of facial expressions dealing the cards at the poker tables it is hard to see that they will eventually make the presence of humans at the table unnecessary.
Until today, no researchers have managed to develop an AI system that reasons just like a human being, although of course, they have been trying to do so. Still one could definitely state that there is much more emphasis on the type of AI that is meant to accomplish tasks and make life for us humans easier. These systems are hardly something to fear since they clearly need human intervention to succeed. It is true that in the process these smart systems terminate certain jobs but at the same time new job opportunities arise so there will still be a need for human manpower despite the robots’ increased abilities.
Getting a Computer to Think for Itself
It might seem impossible to make a computer think for itself. After all, the programmer instructs the computer how to analyze the information it receives. There is a field, though, where scientists are indeed trying to make computers understand and do things without being told so. This field is called genetic programming. This is a method where a population of computer programs is bred in order to better solve problems. The programs are made to transform existing programs into new ones using the Darwinist approach — survival of the fittest — to arrive at the best results. If this sounds a bit confusing, don’t worry. This part of AI is not easy to understand. However, it might come to influence the way we look at computers in the future and make them better at working without our intervention.
Although genetic programming uses terms and expressions from the field of evolution, the programs aren’t actually reproducing themselves. Instead, they are set up to have the ability to select the right program for the problem that needs to be solved. The ability for software to upgrade and select the best possible choice of different options is exactly what is needed to create poker software that can handle more advanced situations where several players are involved in the game and more parameters come into play.
Ruddock Says We Have Nothing to Fear…
In Ruddock’s article we learn that poker players should not fear AI poker. Ruddock means that Cepheus is like Lego Duplo in the world of poker and that it will take an immense amount of time to get anywhere close to Lego or Lego Technic. He admits that Cepheus did beat the poker master Daniel Negreanu in Heads-Up Holdem but he claims that the computer program stands no chance in more complicated situations with several hands and Pot or No Limit rules. His conclusion is that Cepheus can beat bad players but not as fast as a regular good player could. So should we all just relax and forget about AI poker as a concern for fair poker rooms?
Well, Ruddock seem to forget that the majority of players in online poker rooms are indeed very poor players. They don’t necessarily join the big games with many hands. It is exactly this type of player that the computer software knows how to beat and if the software is to be used by poker sites then players will definitely have something to worry about. For a bad player playing against the computer it simply won’t be worth it when the opponent is going to win at all times.
For Learning Purposes
There is however a very important point to Ruddock’s article on AI and poker which has to do with the learning aspect. Cepheus has brought the possibility of making great learning software for players who want to advance faster without losing money on the learning process. Many players are looking for poker schools that let you play over and over again with real situations and scenarios. The advanced computer technique can simulate poker games in a fantastic way, making the practice games exciting and well worth the time spent on them.
As for poker sites using computer software that can outsmart the less skilled players there might be a point to developing certification which assures the players that they are not up against Cepheus but rather regular software where they still have a chance of winning. Perhaps testing facilities like TST (Technical Systems Testing) can incorporate this task of checking whether the poker room is using AI in its poker software or not.
Awareness
There is yet another aspect to consider and that is the awareness of the players themselves. Not many people are aware of the connection between AI and poker and who has actually heard about Cepheus? If people look at AI’s connection to gaming online in the same light as the relationship between online gaming and random number generators popular demand might actually request AI involvement to ensure that the poker games are standardized. Till then, poker rooms can decide to use Cepheus software in games that are mainly used by poor players without ever getting caught. This is an aspect that Ruddock doesn’t bring up in his article. We should all be aware of the fact that AI is something which evolves and that just because the University of Alberta Poker Research Group took well over 20 years to come up with Cepheus it will not necessarily take another 20 to create something even more effective. Ruddock is absolutely right about the fact that the discussion on poker versus AI will continue — and well it should, so that players can become aware of the issue and understand the implications of smart poker software in online poker rooms.
Further reading:
from Darlene Milligan http://ift.tt/1JaffMT via local SEO company
from Tumblr http://ift.tt/1KgpaBT
No comments:
Post a Comment