As part of our ongoing series of monthly focuses on particular areas of technology development, we will in the coming months be looking at the advances being made in artificial intelligence and asking how it might affect the gambling industry. So, in just such a spirit of adventure, we are presenting a couple of stories that you might have missed about how people are already testing AI in gambling-related fields. The first story is about the Swarm AI software platform was used earlier this year to predict the outcome of the Kentucky Derby. The next comes from Sky News and look at how they tested AI to a game of poker.
From TechRepublic, May 2016
Swarm AI predicts the 2016 Kentucky Derby
For those betting on the 142nd Kentucky Derby on Saturday, there are several ways to approach the strategy. Last year, Jimmy Fallon's puppies took a stab at it—and correctly predicted the winner, American Pharoah. Or, you could rely on the experts from the Bleacher Report. Maybe you want to study up on your own, or see which horses are looking good that day.
But there's another method, and it may be better than all the others: Artificial intelligence.
Unanimous AI has built a software platform called UNU that harnesses the power of the crowd to make predictions. Instead of the popular neural network, which emulates brain activity, "swarm AI" looks to a different part of nature: The insect swarm. The concept is that the various units of the group will influence each other in order to arrive at a correct decision—one that is more accurate than any individual prediction.
The tool has been remarkably accurate in the past, predicting winners for the Super Bowl, Oscars, and even predicting how political candidates would perform in the primaries. For the 2015 Oscars, a group of seven UNU users (all non-experts) accurately predicted 11 out of 15 categories correctly—in under a minute. The 73% success rate beat experts at the New York Times, who only achieved 55% accuracy.
I asked Unanimous A.I to do a swarm to predict the Kentucky Derby. On Thursday, they collected a group of 20 participants who claimed to be knowledgeable about the race. The group narrowed the horses down to the top four.
"The UNU swarm intelligence seemed to strongly favor Nyquist for the win, maintaining conviction in that pick even after an unfavorable post position was announced," said Louis Rosenberg, CEO of Unanimous A.I. "Of course it's horse racing, so there's no such thing as a sure thing."
The swarm had met on April 28 before positions were announced and had previously predicted Mor Spirit to be in the top four, but later removed him due to poor post position.
Sky News: Can a computer beat a human at poker? 12 October 2016
Computers have beaten human world champions at chess and, earlier this year, the board game Go. So far, though, they have struggled at the card table. So we challenged one AI to a game.
Why is poker so difficult? Chess and Go are "information complete" games where all players can see all the relevant information. In poker, other players' cards are hidden, making it an "information incomplete" game. Players have to guess opponents' hands from their actions - tricky for computers.
Poker has become a new benchmark for AI research. Solving poker could lead to breakthroughs in all sorts of real life scenarios, from cybersecurity to driverless cars.
Scientists believe it is only a matter of time before AI once again vanquishes humans, hence our human-machine match-up in a game of Texas Hold'Em Limit Poker.
The AI was developed by Johannes Heinrich, researcher studying machine learning at UCL. It combines two techniques: neural networks and reinforcement learning.
Neural networks to some extent mimic the structure of human brains: their processors are highly interconnected and work at the same time to solve problems. They are good at spotting patterns in huge amounts of data.
Reinforcement learning is when a machine, given a task, carries it out, learning from mistakes it makes. In this case, it means playing poker against itself billions of times to get better. Mr Heinrich told Sky News: "Today we are presenting a novel algorithm that has learned in a different way, more similar to how humans learn.
"In particular, it is able to learn abstract patterns, represented by its neural network, that allow it to generalise to new and unseen situations."
This is a particularly beneficial property if we want to apply these methods beyond just poker to larger scale real world applications."
After two hours of quite defensive play, from the computer at least, we called it a draw.
In fact, this particular form of poker - two player, with limits on bets - was last year "weakly solved" in a lab by researchers at the University of Alberta. Its AI would at least break even against a human opponent over the long run.
But more complicated versions, with multiple players and unlimited bets, remain beyond AI. For now.