{"id":1071,"date":"2021-08-31T16:40:48","date_gmt":"2021-08-31T21:40:48","guid":{"rendered":"https:\/\/singularityumexicosummit.com\/?p=1071"},"modified":"2021-08-31T16:40:48","modified_gmt":"2021-08-31T21:40:48","slug":"the-deck-is-not-rigged-poker-and-the-limits-of-ai","status":"publish","type":"post","link":"https:\/\/singularityumexico.com\/en\/the-deck-is-not-rigged-poker-and-the-limits-of-ai\/","title":{"rendered":"The Deck Is Not Rigged: Poker and the Limits of AI"},"content":{"rendered":"[iframe src=”https:\/\/spkt.io\/a\/678541″ width=”100%” height=”100″]\n\n\n\n

Tuomas Sandholm, a computer scientist at Carnegie Mellon University, is not a poker player\u2014or much of a poker fan, in fact\u2014but 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 precisely\u2014a view shared years later by Sandholm in his research with artificial intelligence<\/a>.<\/p>\n\n\n\n

\u201cPoker is the main benchmark and challenge program for games of imperfect information,\u201d 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.<\/p>\n\n\n\n

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<\/a>. (When we met, Libratus was still a toddler and Pluribus didn\u2019t yet exist.) The goal isn\u2019t to solve poker, as such, but to create algorithms whose decision making prowess in poker\u2019s world of imperfect information and stochastic situations\u2014situations that are randomly determined and unable to be predicted\u2014can then be applied to other stochastic realms, like the military, business, government, cybersecurity, even health care.<\/p>\n\n\n\n

While the first program, Claudico, was summarily beaten by human poker players\u2014\u201cone broke-ass robot,\u201d an observer called it\u2014Libratus has triumphed in a series of one-on-one, or heads-up, matches against some of the best online players in the United States.<\/p>\n\n\n\n

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 wasn\u2019t chosen would have yielded a better outcome than one that was. \u201cIntuitively, regret represents how much the AI regrets having not chosen that action in the past,\u201d says Sandholm. The higher the regret, the higher the chance of choosing that action next time.<\/p>\n\n\n\n

It\u2019s a useful way of thinking\u2014but 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, it\u2019s an emotionally laden calculus, and we typically fail to apply it in quite the right way. For a computer, it\u2019s 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?<\/p>\n\n\n\n

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. It\u2019s dangerous to try to exploit your opponent\u2014it opens you up to the risk that you\u2019ll get exploited right back, especially if you\u2019re a computer program and your opponent is human. So instead of attempting to do that, the self-improver lets the opponent\u2019s actions inform the areas where the program should focus. \u201cThat lets the opponent\u2019s actions tell us where [they] think they\u2019ve found holes in our strategy,\u201d Sandholm explained. This allows the algorithm to develop a blueprint strategy to patch those holes.<\/p>\n\n\n\n

It\u2019s a very human-like adaptation, if you think about it. I\u2019m not going to try to outmaneuver you head on. Instead, I\u2019m going to see how you\u2019re trying to outmaneuver me and respond accordingly. Sun-Tzu would surely approve. Watch how you\u2019re perceived, not how you perceive yourself\u2014because in the end, you\u2019re 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.<\/p>\n\n\n\n

There\u2019s one final thing Libratus is able to do: play in situations with unknown probabilities. There\u2019s 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 opponent\u2019s hand trembles, they take a wrong action, and you\u2019re 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, there\u2019s a contingency plan.<\/p>\n\n\n\n

Of course, no algorithm is perfect. When Libratus is playing poker, it\u2019s 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-sum\u2014cyber warfare comes to mind\u2014many 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.<\/p>\n\n\n\n

What\u2019s 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 world\u2019s first kidney exchange. Do you want to be more efficient, giving the maximum number of kidneys as quickly as possible\u2014or more fair, which may come at a cost to efficiency? Do you want as many lives as possible saved\u2014or 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. It\u2019s essential, Sandholm says, to separate means and the ends. To figure out the ends, a human has to decide what the goal is.<\/p>\n\n\n\n

\u201cThe world will ultimately become a lot safer with the help of algorithms like Libratus,\u201d Sandholm told me. I wasn\u2019t 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.<\/p>\n\n\n\n

\u201cLogic is good, and the AI is much better at strategic reasoning than humans can ever be,\u201d he explained. \u201cIt\u2019s taking out irrationality, emotionality. And it\u2019s fairer. If you have an AI on your side, it can lift non-experts to the level of experts. Na\u00efve negotiators will suddenly have a better weapon. We can start to close off the digital divide.\u201d<\/p>\n\n\n\n

It was an optimistic note to end on\u2014a zero-sum, competitive game yielding a more ultimately fair and rational world.<\/p>\n\n\n\n

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 Center\u2014the place that runs all of Sandholm\u2019s poker-AI programs\u2014to the actual processing center that make undertakings like Libratus possible.<\/p>\n\n\n\n

A half-hour drive found us in a parking lot by a large glass building. I\u2019d expected something more futuristic, not the same square, corporate glass squares I\u2019ve seen countless times before. The inside, however, was more promising. First the security checkpoint. Then the ride in the elevator \u2014 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 don\u2019t 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.<\/p>\n\n\n\n

\u201cLet me tell you what you\u2019re going to see before we walk in,\u201d Nystrom told me. \u201cOnce we get inside, it will be too loud to hear.\u201d<\/p>\n\n\n\n

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 \u201ccold\u201d rows alternating with \u201chot\u201d\u2014fans 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.<\/p>\n\n\n\n

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 Warhol\u2019s Pittsburghian origins).<\/p>\n\n\n\n

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?<\/p>\n\n\n\n

Bridges, it turned out, isn\u2019t a single computer. It\u2019s 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. That\u2019s a whole lot of computing power. And that\u2019s only to succeed at heads-up poker, in limited circumstances.<\/p>\n\n\n\n

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 weren\u2019t 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.<\/p>\n\n\n\n

But there\u2019s still much it can\u2019t do: play more opponents, play live, or win every time. There\u2019s more humanity in poker than Libratus has yet conquered. \u201cThere\u2019s this belief that it\u2019s all about statistics and correlations. And we actually don\u2019t believe that,\u201d Nystrom explained as we left Bridges behind. \u201cOnce in a while correlations are good, but in general, they can also be really misleading.\u201d<\/p>\n\n\n\n

Two years later, the Sandholm lab will produce Pluribus. Pluribus will be able to play against five players\u2014and 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.<\/p>\n\n\n\n

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 \u201cthe little tactics of deception, of asking yourself what is the other man going to think I mean to do,\u201d as von Neumann put it?<\/p>\n\n\n\n

Long before I\u2019d spoken to Sandholm, I\u2019d 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 MIT\u2019s Media Lab. Slavin has a decidedly different view from the creators of Pluribus. \u201cOn the one hand, [von Neumann] was a genius,\u201d Kevin Slavin reflects. \u201cBut the presumptuousness of it.\u201d<\/p>\n\n\n\n

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 luck\u2014but 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 \u201cThat they can generate this fabulous, complex randomness.\u201d His opinion of the algorithmic assaults on chance? \u201cThis is their moment,\u201d he said. \u201cBut it\u2019s the exact opposite of what\u2019s really beautiful about a computer, which is that it can do something that\u2019s actually unpredictable. That, to me, is the magic.\u201d<\/p>\n\n\n\n

Will they actually succeed in making the unpredictable predictable, though? That\u2019s what I want to know. Because everything I\u2019ve seen tells me that absolute success is impossible. The deck is not rigged.<\/p>\n\n\n\n

\u201cIt\u2019s an unbelievable amount of work to get there. What do you get at the end? Let\u2019s say they\u2019re successful. Then we live in a world where there\u2019s no God, agency, or luck,\u201d Slavin responded.<\/p>\n\n\n\n

\u201cI don\u2019t want to live there,\u2019\u2019 he added \u201cI just don\u2019t want to live there.\u201d<\/p>\n\n\n\n

Luckily, it seems that for now, he won\u2019t have to. There are more things in life than are yet written in the algorithms. We have no reliable lie detection software\u2014whether 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 can\u2019t get at the reasons for that discomfort: lying, fatigue, stress\u2014they all look much the same. And humans, of course, can also mimic stress where none exists, complicating the picture even further.<\/p>\n\n\n\n

Pluribus may turn out to be powerful, but von Neumann\u2019s challenge still stands: The true nature of games, the most human of the human, remains to be conquered.<\/p>\n\n\n\n

\n\n\n\n

This article was originally published on <\/em>Undark<\/a>. Read the original article<\/a>.<\/em><\/p>\n\n\n\n

Image Credit: Jos\u00e9 Pablo Iglesias<\/a> \/ Unsplash<\/a><\/em><\/p>\n\n\n\n

Author:<\/h4>\n\n\n\n

<\/a>MARIA KONNIKOVA<\/a> Maria Konnikova is the author, most recently, of \u201cThe Biggest Bluff.\u201d She is a regularly contributing writer for The New Yorker, the author of two previous New York Times best-sellers, and a professional poker player. Learn More<\/a><\/p>\n\n\n\n

Original Article<\/a><\/strong><\/p>\n","protected":false},"excerpt":{"rendered":"

Tuomas Sandholm, a computer scientist at Carnegie Mellon University, is not a poker player\u2014or much of a poker fan, in fact\u2014but 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 […]\n","protected":false},"author":1,"featured_media":1072,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"episode_type":"","audio_file":"","podmotor_file_id":"","podmotor_episode_id":"","cover_image":"","cover_image_id":"","duration":"","filesize":"","filesize_raw":"","date_recorded":"","explicit":"","block":"","footnotes":""},"categories":[13],"tags":[],"series":[],"class_list":["post-1071","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-articulos-ingles"],"episode_featured_image":"https:\/\/singularityumexico.com\/wp-content\/uploads\/2021\/08\/red-deck-of-cards-ace-of-spades-1068x601-1.jpg","episode_player_image":"https:\/\/singularityumexico.com\/wp-content\/uploads\/2023\/05\/11711533-1673157178559-89a95be153719-4-scaled.jpg","download_link":"","player_link":"","audio_player":false,"episode_data":{"playerMode":"dark","subscribeUrls":{"apple_podcasts":{"key":"apple_podcasts","url":"","label":"Apple Podcasts","class":"apple_podcasts","icon":"apple-podcasts.png"},"stitcher":{"key":"stitcher","url":"","label":"Stitcher","class":"stitcher","icon":"stitcher.png"},"google_podcasts":{"key":"google_podcasts","url":"","label":"Google Podcasts","class":"google_podcasts","icon":"google-podcasts.png"},"spotify":{"key":"spotify","url":"","label":"Spotify","class":"spotify","icon":"spotify.png"}},"rssFeedUrl":"https:\/\/singularityumexico.com\/en\/feed\/podcast\/the-feedback-loop-by-singularity","embedCode":"

The Deck Is Not Rigged: Poker and the Limits of AI<\/a><\/blockquote>