Tuomas Sandholm, a computer scientist at Carnegie Mellon University, is not a poker playeror much of a poker fan, in factbut he is fascinated by the game for much the same reason as the great game theorist John von Neumann before him. Von Neumann, who died in 1957, viewed poker as the perfect model for human decision making, for finding the balance between skill and chance that accompanies our every choice. He saw poker as the ultimate strategic challenge, combining as it does not just the mathematical elements of a game like chess but the uniquely human, psychological angles that are more difficult to model preciselya view shared years later by Sandholm in his research with artificial intelligence.
Poker is the main benchmark and challenge program for games of imperfect information, Sandholm told me on a warm spring afternoon in 2018, when we met in his offices in Pittsburgh. The game, it turns out, has become the gold standard for developing artificial intelligence.
Tall and thin, with wire-frame glasses and neat brow hair framing a friendly face, Sandholm is behind the creation of three computer programs designed to test their mettle against human poker players: Claudico, Libratus, and most recently, Pluribus. (When we met, Libratus was still a toddler and Pluribus didnt yet exist.) The goal isnt to solve poker, as such, but to create algorithms whose decision making prowess in pokers world of imperfect information and stochastic situationssituations that are randomly determined and unable to be predictedcan then be applied to other stochastic realms, like the military, business, government, cybersecurity, even health care.
While the first program, Claudico, was summarily beaten by human poker playersone broke-ass robot, an observer called itLibratus has triumphed in a series of one-on-one, or heads-up, matches against some of the best online players in the United States.
Libratus relies on three main modules. The first involves a basic blueprint strategy for the whole game, allowing it to reach a much faster equilibrium than its predecessor. It includes an algorithm called the Monte Carlo Counterfactual Regret Minimization, which evaluates all future actions to figure out which one would cause the least amount of regret. Regret, of course, is a human emotion. Regret for a computer simply means realizing that an action that wasnt chosen would have yielded a better outcome than one that was. Intuitively, regret represents how much the AI regrets having not chosen that action in the past, says Sandholm. The higher the regret, the higher the chance of choosing that action next time.
Its a useful way of thinkingbut one that is incredibly difficult for the human mind to implement. We are notoriously bad at anticipating our future emotions. How much will we regret doing something? How much will we regret not doing something else? For us, its an emotionally laden calculus, and we typically fail to apply it in quite the right way. For a computer, its all about the computation of values. What does it regret not doing the most, the thing that would have yielded the highest possible expected value?
The second module is a sub-game solver that takes into account the mistakes the opponent has made so far and accounts for every hand she could possibly have. And finally, there is a self-improver. This is the area where data and machine learning come into play. Its dangerous to try to exploit your opponentit opens you up to the risk that youll get exploited right back, especially if youre a computer program and your opponent is human. So instead of attempting to do that, the self-improver lets the opponents actions inform the areas where the program should focus. That lets the opponents actions tell us where [they] think theyve found holes in our strategy, Sandholm explained. This allows the algorithm to develop a blueprint strategy to patch those holes.
Its a very human-like adaptation, if you think about it. Im not going to try to outmaneuver you head on. Instead, Im going to see how youre trying to outmaneuver me and respond accordingly. Sun-Tzu would surely approve. Watch how youre perceived, not how you perceive yourselfbecause in the end, youre playing against those who are doing the perceiving, and their opinion, right or not, is the only one that matters when you craft your strategy. Overnight, the algorithm patches up its overall approach according to the resulting analysis.
Theres one final thing Libratus is able to do: play in situations with unknown probabilities. Theres a concept in game theory known as the trembling hand: There are branches of the game tree that, under an optimal strategy, one should theoretically never get to; but with some probability, your all-too-human opponents hand trembles, they take a wrong action, and youre suddenly in a totally unmapped part of the game. Before, that would spell disaster for the computer: An unmapped part of the tree means the program no longer knows how to respond. Now, theres a contingency plan.
Of course, no algorithm is perfect. When Libratus is playing poker, its essentially working in a zero-sum environment. It wins, the opponent loses. The opponent wins, it loses. But while some real-life interactions really are zero-sumcyber warfare comes to mindmany others are not nearly as straightforward: My win does not necessarily mean your loss. The pie is not fixed, and our interactions may be more positive-sum than not.
Whats more, real-life applications have to contend with something that a poker algorithm does not: the weights that are assigned to different elements of a decision. In poker, this is a simple value-maximizing process. But what is value in the human realm? Sandholm had to contend with this before, when he helped craft the worlds first kidney exchange. Do you want to be more efficient, giving the maximum number of kidneys as quickly as possibleor more fair, which may come at a cost to efficiency? Do you want as many lives as possible savedor do some take priority at the cost of reaching more? Is there a preference for the length of the wait until a transplant? Do kids get preference? And on and on. Its essential, Sandholm says, to separate means and the ends. To figure out the ends, a human has to decide what the goal is.
The world will ultimately become a lot safer with the help of algorithms like Libratus, Sandholm told me. I wasnt sure what he meant. The last thing that most people would do is call poker, with its competition, its winners and losers, its quest to gain the maximum edge over your opponent, a haven of safety.
Logic is good, and the AI is much better at strategic reasoning than humans can ever be, he explained. Its taking out irrationality, emotionality. And its fairer. If you have an AI on your side, it can lift non-experts to the level of experts. Nave negotiators will suddenly have a better weapon. We can start to close off the digital divide.
It was an optimistic note to end ona zero-sum, competitive game yielding a more ultimately fair and rational world.
I wanted to learn more, to see if it was really possible that mathematics and algorithms could ultimately be the future of more human, more psychological interactions. And so, later that day, I accompanied Nick Nystrom, the chief scientist of the Pittsburgh Supercomputing Centerthe place that runs all of Sandholms poker-AI programsto the actual processing center that make undertakings like Libratus possible.
A half-hour drive found us in a parking lot by a large glass building. Id expected something more futuristic, not the same square, corporate glass squares Ive seen countless times before. The inside, however, was more promising. First the security checkpoint. Then the ride in the elevator down, not up, to roughly three stories below ground, where we found ourselves in a maze of corridors with card readers at every juncture to make sure you dont slip through undetected. A red-lit panel formed the final barrier, leading to a small sliver of space between two sets of doors. I could hear a loud hum coming from the far side.
Let me tell you what youre going to see before we walk in, Nystrom told me. Once we get inside, it will be too loud to hear.
I was about to witness the heart of the supercomputing center: 27 large containers, in neat rows, each housing multiple processors with speeds and abilities too great for my mind to wrap around. Inside, the temperature is by turns arctic and tropic, so-called cold rows alternating with hotfans operate around the clock to cool the processors as they churn through millions of giga, mega, tera, peta and other ever-increasing scales of data bytes. In the cool rows, robotic-looking lights blink green and blue in orderly progression. In the hot rows, a jumble of multicolored wires crisscrosses in tangled skeins.
In the corners stood machines that had outlived their heyday. There was Sherlock, an old Cray model, that warmed my heart. There was a sad nameless computer, whose anonymity was partially compensated for by the Warhol soup cans adorning its cage (an homage to Warhols Pittsburghian origins).
And where does Libratus live, I asked? Which of these computers is Bridges, the computer that runs the AI Sandholm and I had been discussing?
Bridges, it turned out, isnt a single computer. Its a system with processing power beyond comprehension. It takes over two and a half petabytes to run Libratus. A single petabyte is a million gigabytes: You could watch over 13 years of HD video, store 10 billion photos, catalog the contents of the entire Library of Congress word for word. Thats a whole lot of computing power. And thats only to succeed at heads-up poker, in limited circumstances.
Yet despite the breathtaking computing power at its disposal, Libratus is still severely limited. Yes, it beat its opponents where Claudico failed. But the poker professionals werent allowed to use many of the tools of their trade, including the opponent analysis software that they depend on in actual online games. And humans tire. Libratus can churn for a two-week marathon, where the human mind falters.
But theres still much it cant do: play more opponents, play live, or win every time. Theres more humanity in poker than Libratus has yet conquered. Theres this belief that its all about statistics and correlations. And we actually dont believe that, Nystrom explained as we left Bridges behind. Once in a while correlations are good, but in general, they can also be really misleading.
Two years later, the Sandholm lab will produce Pluribus. Pluribus will be able to play against five playersand will run on a single computer. Much of the human edge will have evaporated in a short, very short time. The algorithms have improved, as have the computers. AI, it seems, has gained by leaps and bounds.
So does that mean that, ultimately, the algorithmic can indeed beat out the human, that computation can untangle the web of human interaction by discerning the little tactics of deception, of asking yourself what is the other man going to think I mean to do, as von Neumann put it?
Long before Id spoken to Sandholm, Id met Kevin Slavin, a polymath of sorts whose past careers have including founding a game design company and an interactive art space and launching the Playful Systems group at MITs Media Lab. Slavin has a decidedly different view from the creators of Pluribus. On the one hand, [von Neumann] was a genius, Kevin Slavin reflects. But the presumptuousness of it.
Slavin is firmly on the side of the gambler, who recognizes uncertainty for what it is and thus is able to take calculated risks when necessary, all the while tampering confidence at the outcome. The most you can do is put yourself in the path of luckbut to think you can guess with certainty the actual outcome is a presumptuousness the true poker player foregoes. For Slavin, the wonder of computers is That they can generate this fabulous, complex randomness. His opinion of the algorithmic assaults on chance? This is their moment, he said. But its the exact opposite of whats really beautiful about a computer, which is that it can do something thats actually unpredictable. That, to me, is the magic.
Will they actually succeed in making the unpredictable predictable, though? Thats what I want to know. Because everything Ive seen tells me that absolute success is impossible. The deck is not rigged.
Its an unbelievable amount of work to get there. What do you get at the end? Lets say theyre successful. Then we live in a world where theres no God, agency, or luck, Slavin responded.
I dont want to live there, he added I just dont want to live there.
Luckily, it seems that for now, he wont have to. There are more things in life than are yet written in the algorithms. We have no reliable lie detection softwarewhether in the face, the skin, or the brain. In a recent test of bluffing in poker, computer face recognition failed miserably. We can get at discomfort, but we cant get at the reasons for that discomfort: lying, fatigue, stressthey all look much the same. And humans, of course, can also mimic stress where none exists, complicating the picture even further.
Pluribus may turn out to be powerful, but von Neumanns challenge still stands: The true nature of games, the most human of the human, remains to be conquered.
This article was originally published on Undark. Read the original article.
Image Credit: Jos Pablo Iglesias /Unsplash
- JUST LAUNCHED Interactive Texas Hold'em Poker Quiz: Which hand wins? - European Gaming Industry News - June 5th, 2020 [June 5th, 2020]
- The hilarious story of how Phil Mickelson rigged a hand of poker against his brother - Golf Digest - June 5th, 2020 [June 5th, 2020]
- US Live Poker Rooms Begin To Reopen With Major Changes - Legal Gambling News - June 5th, 2020 [June 5th, 2020]
- Loto-Quebec hit with class action suit over exposed online poker hole cards - CalvinAyre.com - June 5th, 2020 [June 5th, 2020]
- Who was Poker Alice? The gunslinging gambler of the Wild West who may have come from Sudbury - East Anglian Daily Times - June 9th, 2020 [June 9th, 2020]
- Odds of Being Dealt a Royal Flush - Royal Flush Odds and More - BestUSCasinos.org - June 9th, 2020 [June 9th, 2020]
- David Haye Dreaming of Being a Poker World Champion - PokerTube - June 9th, 2020 [June 9th, 2020]
- Avoid the Sharks by Playing Stream Satellites on Unibet Poker - PokerNews.com - June 9th, 2020 [June 9th, 2020]
- Bans on smoking and poker are part of plan to reopen Detroit casinos - The Detroit News - June 9th, 2020 [June 9th, 2020]
- Poker Strategy With Jonathan Little: Folding Top Pair, Top Kicker On The Flop - Poker News - CardPlayer.com - June 9th, 2020 [June 9th, 2020]
- Genting Ceases All Live Poker Activity in the United Kingdom - PokerNews.com - June 21st, 2020 [June 21st, 2020]
- The best actors to have appeared around a poker table - CineVue - June 21st, 2020 [June 21st, 2020]
- Ike Haxton Raises Concerns About the Health of Live Poker Players - HighstakesDB - June 21st, 2020 [June 21st, 2020]
- Book Review - Poker... What does gambling have to do with it? - Books - Ahram Online - June 21st, 2020 [June 21st, 2020]
- Will the online poker boom last? - Casino Beats - June 21st, 2020 [June 21st, 2020]
- Online poker 'booming' in lockdown at every level of the game - theScore - June 21st, 2020 [June 21st, 2020]
- Live Poker Tournaments Are Slowly Returning Around The Globe - Poker News - CardPlayer.com - June 21st, 2020 [June 21st, 2020]
- How a psychologist cracked the secret of winning at professional poker in just a year - Telegraph.co.uk - June 21st, 2020 [June 21st, 2020]
- MASTERING THE GAME OF POKER WITH MACHINE LEARNING: How Machine Learning Impacts the World of Online Poker - The Fan Carpet - June 21st, 2020 [June 21st, 2020]
- Poker Worlds Lost Summer: What Happens When Live-Action Sharks Are Forced to Stay Home? - Deadspin - June 21st, 2020 [June 21st, 2020]
- Late Night Poker: A Week That Changed The World - Padraig Parkinson Blog - CardPlayer.com - June 21st, 2020 [June 21st, 2020]
- 'It Would Have Changed the Game' -- An Oral History of ChipTic, Part 2 - PokerNews.com - July 10th, 2020 [July 10th, 2020]
- Bill filed to expand live poker in New Jersey - Online Poker.net - July 10th, 2020 [July 10th, 2020]
- Battle for India Online Poker Championship 2020: Who will win the crown? - Sportskeeda - July 10th, 2020 [July 10th, 2020]
- GGPoker's $5 Million WSOP Season Giveaway is the Biggest in the History of Online Poker - Pokerfuse - July 10th, 2020 [July 10th, 2020]
- Canterbury returns to 24/7 hours of operation, poker is back today - SW News Media - July 10th, 2020 [July 10th, 2020]
- The winning Lotto Poker numbers and results for Friday July 10, 2020 are in... - Born2Invest - July 11th, 2020 [July 11th, 2020]
- Casinos and Poker Rooms open, despite rise in Covid-19 casesciting measures to safely operate. - Northeast Valley News - July 11th, 2020 [July 11th, 2020]
- Sunflower Diversified hosts annual poker run; seeks sponsorships - Great Bend Tribune - July 11th, 2020 [July 11th, 2020]
- The winning Poker Lotto numbers and results for Saturday July 11, 2020 are in... - Born2Invest - July 11th, 2020 [July 11th, 2020]
- Join the Newly-Launched PokerNews Discord Server Today! - PokerNews.com - July 11th, 2020 [July 11th, 2020]
- Looking to Join the PA Online Poker Party: partypoker Nearing Regulatory Approval - Play Pennsylvania - July 11th, 2020 [July 11th, 2020]
- The Rise and Fall of Dani Stern - HighstakesDB - July 12th, 2020 [July 12th, 2020]
- Pat McAfee Won $1,400 in an Underground Poker Game and Used It to Make His NFL Dream Come True - Sportscasting - July 12th, 2020 [July 12th, 2020]
- Global Video Poker Machines Market Analysis and Forecast 2027- including drivers, constraints, intimidation, challenges, opportunities, and... - July 13th, 2020 [July 13th, 2020]
- Panicking Professional Poker Player Suddenly Cant Remember Whether Ace Is Better Than King - The Onion - July 13th, 2020 [July 13th, 2020]
- Dennis Atkins: The stakes couldnt be higher in PMs poker game with China - The New Daily - July 13th, 2020 [July 13th, 2020]
- Kami Hudson Finds Satisfaction In Poker, Happiness In Honduras - PocketFives - July 13th, 2020 [July 13th, 2020]
- Seven Types of Bets You Can Make In Poker - Poker News - CardPlayer.com - July 13th, 2020 [July 13th, 2020]
- Transitioning From Live Poker To Online Poker | My Blog - Tight Poker - July 14th, 2020 [July 14th, 2020]
- Poker Market Size By Product Analysis, Application, End-Users, Regional Outlook, Competitive Strategies And Forecast Up To 2026 - 3rd Watch News - July 16th, 2020 [July 16th, 2020]
- PhilLaak the other one wins 1st WSOP bracelet - Las Vegas Review-Journal - July 16th, 2020 [July 16th, 2020]
- PokerStars To Host Online Poker Events for the Mind Sports Olympiad - Pokerfuse - July 17th, 2020 [July 17th, 2020]
- Focusing on flushes and straights in video poker - Atlantic City Weekly - July 17th, 2020 [July 17th, 2020]
- Arjun Rampal is playing poker to support NGOs - The Hindu - July 18th, 2020 [July 18th, 2020]
- I won 2 million playing poker - this is what I did next - Hull Live - July 18th, 2020 [July 18th, 2020]
- Advt: 5 reasons why everyone is hooked to online poker! Have you joined the party yet? - Times of India - July 21st, 2020 [July 21st, 2020]
- IGT cleared to debut its Cobalt 27 VLT with new Reel Poker content in Louisiana - Yogonet International - July 21st, 2020 [July 21st, 2020]
- Poker - Wikipedia - July 21st, 2020 [July 21st, 2020]
- Natural8 2020 WSOP Online Hand of the Week: A Wholesome Hand for Ryan Depaulo - PokerNews.com - July 21st, 2020 [July 21st, 2020]
- Flag football shifts to online poker, auction to honour Corey Trudeau - The Kingston Whig-Standard - July 21st, 2020 [July 21st, 2020]
- How Coronavirus changed the perception of Poker - CalvinAyre.com - July 21st, 2020 [July 21st, 2020]
- Orange City OKs more poker players, and Greyhound racing in Daytona is over - The West Volusia Beacon - July 21st, 2020 [July 21st, 2020]
- The Most Famous Woman in the History of Gambling - Poker Alice Story - Newswire - July 21st, 2020 [July 21st, 2020]
- How Has Poker Changed Over Many Years? - Loop21 - July 21st, 2020 [July 21st, 2020]
- Trump's approach to politics bears the hallmarks of a bad poker player, author says - CBC.ca - July 21st, 2020 [July 21st, 2020]
- Key Points To Consider When Choosing Online Poker Tables - Top10PokerWebsites - July 21st, 2020 [July 21st, 2020]
- The Deck Is Not Rigged: Poker and the Limits of AI - Undark Magazine - July 21st, 2020 [July 21st, 2020]
- NJ Special Legislation and the Impact on Online Casino Players in NY - Blog - The Island Now - July 21st, 2020 [July 21st, 2020]
- The Phil Ivey Of Poker - HBCU Buzz - July 21st, 2020 [July 21st, 2020]
- Venetian poker room going to 8-handed tables - Las Vegas Review-Journal - July 21st, 2020 [July 21st, 2020]
- Chasing Poker Greatness: #68 Andrew Brokos: Co-Host ThinkingPoker Podcast & Author of Play Optimal Poker - Pokerfuse - July 21st, 2020 [July 21st, 2020]
- Dont Hold Your Breath: Live Poker Not A Top Priority In Reopened PA Casinos Yet - Penn Bets - July 21st, 2020 [July 21st, 2020]
- Inside the world of Daniel Negreanu: The life and making of poker's biggest superstar - Mirror Online - July 26th, 2020 [July 26th, 2020]
- Best 9 Online Poker Sites for Real Money 2020 - Bonus to ... - July 26th, 2020 [July 26th, 2020]
- Josh & Heather Altman Threw Son Ace an Adorable Poker-Themed 1st Birthday Bash - Bravo - July 26th, 2020 [July 26th, 2020]
- Best 7 Online Poker Sites (2020) for Real Money - July 26th, 2020 [July 26th, 2020]
- 5 Best Online Poker Sites for Real Money (2020) - July 26th, 2020 [July 26th, 2020]
- This book is the new 'Moneyball' (and every investor should read it) - The Australian Financial Review - July 28th, 2020 [July 28th, 2020]
- Man Utd in transfer 'poker game' for winger that could save Ed Woodward 55m - Express - July 29th, 2020 [July 29th, 2020]
- What Are the Best Poker Apps of 2020? - GameIndustry.com - July 30th, 2020 [July 30th, 2020]
- Kai Havertz WILL face Rangers as Chelsea 'poker game' forces dramatic U-turn over 80m superstar - Daily Record - July 30th, 2020 [July 30th, 2020]
- Poker, Metacognition, and 'The Biggest Bluff' | Learning Innovation - Inside Higher Ed - July 30th, 2020 [July 30th, 2020]
- How to attract more luck into your life, using the 'poker mindset' - Business Insider - Business Insider - July 30th, 2020 [July 30th, 2020]
- Facebook AI ReBel Capable of Beating Poker Pros | My Blog - Tight Poker - July 30th, 2020 [July 30th, 2020]
- Hugh Freyer, Poker Player With the Discretion of a Banker, Dies at 86 - The New York Times - July 30th, 2020 [July 30th, 2020]
- Poker, Metacognition and 'The Biggest Bluff' | Learning Innovation - Inside Higher Ed - July 30th, 2020 [July 30th, 2020]
- Facebooks New ReBel Poker AI Scores Higher than Pluribus - Beat The Fish - July 31st, 2020 [July 31st, 2020]
- Ivonne Montealegre discusses how live poker is embracing the online space - CalvinAyre.com - July 31st, 2020 [July 31st, 2020]
- Ethan Yau Brings Audience Into His World Series of Poker Dream - PocketFives - August 1st, 2020 [August 1st, 2020]