Page 1 of 2
1
2
LastLast
  1. #1
    Scarab Lord downnola's Avatar
    10+ Year Old Account
    Join Date
    Feb 2011
    Location
    Made in Philly, living in Akron.
    Posts
    4,572

    Latest A.I. can work things out without being taught.

    https://www.economist.com/news/scien...gs-out-without

    The latest AI can work things out without being taught

    Learning to play Go is only the start

    IN 2016 Lee Sedol, one of the world’s best players of Go, lost a match in Seoul to a computer program called AlphaGo by four games to one. It was a big event, both in the history of Go and in the history of artificial intelligence (AI). Go occupies roughly the same place in the culture of China, Korea and Japan as chess does in the West. After its victory over Mr Lee, AlphaGo beat dozens of renowned human players in a series of anonymous games played online, before re-emerging in May to face Ke Jie, the game’s best player, in Wuzhen, China. Mr Ke fared no better than Mr Lee, losing to the computer 3-0.

    For AI researchers, Go is equally exalted. Chess fell to the machines in 1997, when Garry Kasparov lost a match to Deep Blue, an IBM computer. But until Mr Lee’s defeat, Go’s complexity had made it resistant to the march of machinery. AlphaGo’s victory was an eye-catching demonstration of the power of a type of AI called machine learning, which aims to get computers to teach complicated tasks to themselves.

    AlphaGo learned to play Go by studying thousands of games between expert human opponents, extracting rules and strategies from those games and then refining them in millions more matches which the program played against itself. That was enough to make it stronger than any human player. But researchers at DeepMind, the firm that built AlphaGo, were confident that they could improve it. In a paper just published in Nature they have unveiled the latest version, dubbed AlphaGo Zero. It is much better at the game, learns to play much more quickly and requires far less computing hardware to do well. Most important, though, unlike the original version, AlphaGo Zero has managed to teach itself the game without recourse to human experts at all.

    The eyes have it
    Like all the best games, Go is easy to learn but hard to master. Two players, Black and White, take turns placing stones on the intersections of a board consisting of 19 vertical lines and 19 horizontal ones. The aim is to control more territory than your opponent. Stones that are surrounded by an opponent’s are removed from the board. Players carry on until neither wishes to continue. Each then adds the number of his stones on the board to the number of empty grid intersections he has surrounded. The larger total is the winner.

    The difficulty comes from the sheer number of possible moves. A 19x19 board offers 361 different places on which Black can put the initial stone. White then has 360 options in response, and so on. The total number of legal board arrangements is in the order of 10170, a number so large it defies any physical analogy (there are reckoned to be about 1080 atoms in the observable universe, for instance).
    Human experts focus instead on understanding the game at a higher level. Go’s simple rules give rise to plenty of emergent structure. Players talk of features such as “eyes” and “ladders”, and of concepts such as “threat” and “life-and-death”. But although human players understand such concepts, explaining them in the hyper-literal way needed to program a computer is much harder. Instead, the original AlphaGo studied thousands of examples of human games, a process called supervised learning. Since human play reflects human understanding of such concepts, a computer exposed to enough of it can come to understand those concepts as well. Once AlphaGo had arrived at a decent grasp of tactics and strategy with the help of its human teachers, it kicked away its crutches and began playing millions of unsupervised training games against itself, improving its play with every game.

    Supervised learning is useful for much more than Go. It is the basic idea behind many of the recent advances in AI, helping computers learn to do things such as identify faces in pictures, recognise human speech reliably, filter spam from e-mail efficiently and more. But as Demis Hassabis, Deepmind’s boss, observes, supervised learning has limits. It relies on the availability of training data to feed to the computer to show the machine what it is meant to be doing. Such data must be filtered by human experts. The training data for face recognition, for instance, consist of thousands of pictures, some with faces and some without, each labelled as such by a person. That makes such data sets expensive, assuming they are available at all. And, as the paper points out, there can be more subtle problems. Relying on human experts for guidance risks imposing human limits on a computer’s ability.

    AlphaGo Zero is designed to avoid all these problems by skipping the training-wheels phase entirely. The program starts only with the rules of the game and a “reward function”, which awards it a point for a win and docks a point for a loss. It is then encouraged to experiment, repeatedly playing games against other versions of itself, subject only to the constraint that it must try to maximise its reward by winning as much as possible.

    The program started by placing stones randomly, with no real idea of what it was doing. But it improved rapidly. After a single day it was playing at the level of an advanced professional. After two days it had surpassed the performance of the version that beat Mr Lee in 2016.

    DeepMind’s researchers were able to watch their creation rediscover the Go knowledge that human beings have accumulated over thousands of years. Sometimes, it seemed eerily human-like. After about three hours of training the program was preoccupied with the idea of greedily capturing stones, a phase that most human beginners also go through. At others it seemed decidedly alien. For example, ladders are patterns of stones that extend in a diagonal slash across the board as one player attempts to capture a group of his opponent’s stones. They are frequent features of Go games. Because a ladder consists of a simple, repeating pattern, human novices quickly learn to extrapolate them and work out if building a particular ladder will succeed or fail. But AlphaGo Zero—which is not capable of extrapolation, and instead experiments with new moves semi-randomly—took longer than expected to come to grips with the concept.

    Climbing the ladder
    Nevertheless, learning for itself rather than relying on hints from people seemed, on balance, to be a big advantage. For example, joseki are specialised sequences of well-known moves that take place near the edges of the board. (Their scripted nature makes them a little like chess openings.) AlphaGo Zero discovered the standard joseki taught to human players. But it also discovered, and eventually preferred, several others that were entirely of its own invention. The machine, says David Silver, who led the AlphaGo project, seemed to play with a distinctly non-human style.
    The result is a program that is not just superhuman, but crushingly so. Skill at Go (and chess, and many other games) can be quantified with something called an Elo rating, which gives the probability, based on past performance, that one player will beat another. A player has a 50:50 chance of beating an opponent with the same Elo rating, but only a 25% chance of beating one with a rating 200 points higher. Mr Ke has a rating of 3,661. Mr Lee’s is 3,526. After 40 days of training AlphaGo Zero had an Elo rating of more than 5,000—putting it as far ahead of Mr Ke as Mr Ke is of a keen amateur, and suggesting that it is, in practice, impossible for Mr Ke, or any other human being, ever to defeat it. When it played against the version of AlphaGo that first beat Mr Lee, it won by 100 games to zero.

    There is, of course, more to life than Go. Algorithms such as the ones that power the various iterations of AlphaGo might, its creators hope, be applied to other tasks that are conceptually similar. (DeepMind has already used those that underlie the original AlphaGo to help Google slash the power consumption of its data centres.) But an algorithm that can learn without guidance from people means that machines can be let loose on problems that people do not understand how to solve. Anything that boils down to an intelligent search through an enormous number of possibilities, said Mr Hassabis, could benefit from AlphaGo’s approach. He cited classic thorny problems such as working out how proteins fold into their final, functional shapes, predicting which molecules might have promise as medicines, or accurately simulating chemical reactions.

    Advances in AI often trigger worries about human obsolescence. DeepMind hopes such machines will end up as assistants to biological brains, rather than replacements for them, in the way that other technologies from search engines to paper have done. Watching a machine invent new ways to tackle a problem can, after all, help push people down new and productive paths. One of the benefits of AlphaGo, says Mr Silver, is that, in a game full of history and tradition, it has encouraged human players to question the old wisdom, and to experiment. After losing to AlphaGo, Mr Ke studied the computer’s moves, looking for ideas. He then went on a 22-game winning streak against human opponents, an impressive feat even for someone of his skill. Supervised learning, after all, can work in both directions.
    I'd be interested in seeing A.I. tackle a complex card game like Poker or Magic the Gathering. Go players were amazed at how familiar, but alien the A.I. strategy is. I could easily see MTG's Modern format looking completely different if A.I. got a hold of the card database and did it's thing.
    Populists (and "national socialists") look at the supposedly secret deals that run the world "behind the scenes". Child's play. Except that childishness is sinister in adults.
    - Christopher Hitchens

  2. #2
    Dreadlord Gadion's Avatar
    10+ Year Old Account
    Join Date
    Jun 2013
    Location
    I Live On The Web
    Posts
    842
    Quote Originally Posted by downnola View Post
    https://www.economist.com/news/scien...gs-out-without



    I'd be interested in seeing A.I. tackle a complex card game like Poker or Magic the Gathering. Go players were amazed at how familiar, but alien the A.I. strategy is. I could easily see MTG's Modern format looking completely different if A.I. got a hold of the card database and did it's thing.
    The way I see it, any game with a clear set of rules has limitations and the sheer processing power of AI makes it superior to humans. Playing against an AI is pointless, because at some point it will become impossible to win. Admittedly, this is less so in games of chance, but they have better abilities of calculating the statistical odds.

    I'd be more interested in real life applications that have less defined rule sets. Once an AI that knows nothing gets to learn anything by itself without guidance, that would be the birth of the AI species' dominance on Earth

  3. #3

  4. #4
    I was born a hundred, two hundred years early.
    .

    "This will be a fight against overwhelming odds from which survival cannot be expected. We will do what damage we can."

    -- Capt. Copeland

  5. #5
    Quote Originally Posted by Gadion View Post
    The way I see it, any game with a clear set of rules has limitations and the sheer processing power of AI makes it superior to humans. Playing against an AI is pointless, because at some point it will become impossible to win. Admittedly, this is less so in games of chance, but they have better abilities of calculating the statistical odds.

    I'd be more interested in real life applications that have less defined rule sets. Once an AI that knows nothing gets to learn anything by itself without guidance, that would be the birth of the AI species' dominance on Earth
    It depends on the randomness involved as well. You can plan for an outcome and pick the higher chance something occurs but there is a chance something else happens. Like in Cards or dice.

    Its why games like checkers and tic-tack-toe can be "solved" there's no real randomness to playing the best way.

    In poker, backgammon, axis and allies the random get involved.

  6. #6

  7. #7
    Scarab Lord downnola's Avatar
    10+ Year Old Account
    Join Date
    Feb 2011
    Location
    Made in Philly, living in Akron.
    Posts
    4,572
    I would pay good money to have someone sneak "gg ez scrub" into the algorithm.
    Populists (and "national socialists") look at the supposedly secret deals that run the world "behind the scenes". Child's play. Except that childishness is sinister in adults.
    - Christopher Hitchens

  8. #8
    Dreadlord Gadion's Avatar
    10+ Year Old Account
    Join Date
    Jun 2013
    Location
    I Live On The Web
    Posts
    842
    Quote Originally Posted by Logwyn View Post
    It depends on the randomness involved as well. You can plan for an outcome and pick the higher chance something occurs but there is a chance something else happens. Like in Cards or dice.

    Its why games like checkers and tic-tack-toe can be "solved" there's no real randomness to playing the best way.

    In poker, backgammon, axis and allies the random get involved.
    As I said, their dominance is lessened in games of chance. However perform the games a great enough amount of times and there will probably be a better than even chance that the computer is the winner. They simply have better information and are likely to make better decisions. If they are programmed to be overly cautious though they may fold or neglect to play and then the human opponent may win by default.

  9. #9
    I Don't Work Here Endus's Avatar
    10+ Year Old Account
    Join Date
    Feb 2010
    Location
    Ottawa, ON
    Posts
    78,905
    Quote Originally Posted by downnola View Post
    I'd be interested in seeing A.I. tackle a complex card game like Poker or Magic the Gathering. Go players were amazed at how familiar, but alien the A.I. strategy is. I could easily see MTG's Modern format looking completely different if A.I. got a hold of the card database and did it's thing.
    With something like poker, the issue is that you're not just playing by the rules of the game; the whole betting angle is based about 20% on understanding odds, and the remaining 80% is in being able to read other people. Because winning, in the long run, isn't about "who gets the best hands the most often", it's about who most consistently bets big and wins, and cuts losses early when their chances are slim. And to do that, you need to be able to figure out if your buddy is bluffing you into folding when he's got squat, or if he's sitting on a kickass hand.

    MtG would be really complex to design, but a computer is likely to outperform a human player once that complexity's built in. Any strategic game, a computer is better going to be able to establish the odds of any given play and keep track of the broad list of possible responses.

    A good example of this is how kickass the Watson AI was on Jeopardy. Understanding language is a complex problem, and while it gloriously fucked it up occasionally, it was more consistently able to get the answer right than the best human players. They put that system up against the two best human players Jeopardy had ever seen, and it crushed them. And it would've been even more crushing if the AI could properly understand language; some tricky question types that humans can understand fine tripped it up due to the lack of syntax.


  10. #10
    The Unstoppable Force PC2's Avatar
    7+ Year Old Account
    Join Date
    Feb 2015
    Location
    California
    Posts
    21,877
    Libratus from earlier this year already showed AI can beat poker pros. Even bluffing has a rhyme and rhythm that can be exploited.

    For Alpha Go Zero the big idea was that they can turn brute force tree search into a minimal policy improver. Where as in previous programs like Deep Blue and Alpha Go vanilla, tree search was fundamentally baked in to how it works.

  11. #11
    The Undying
    15+ Year Old Account
    Join Date
    Aug 2007
    Location
    the Quiet Room
    Posts
    34,524
    Quote Originally Posted by downnola View Post
    https://www.economist.com/news/scien...gs-out-without



    I'd be interested in seeing A.I. tackle a complex card game like Poker or Magic the Gathering. Go players were amazed at how familiar, but alien the A.I. strategy is. I could easily see MTG's Modern format looking completely different if A.I. got a hold of the card database and did it's thing.
    Very fascinating - thanks for linking it and posting.

    I find the idea of A.I. taking over the world and crushing humans interesting. Obviously, if that happened, it would be bad. My biggest question in all of it is where coding can limit what an A.I. does - and therefore prevent it from actually taking over the world, even if it runs most of it in the near-future.

    - - - Updated - - -

    Quote Originally Posted by Hubcap View Post
    I was born a hundred, two hundred years early.
    Maybe not. We might be the generation that is at the cusp of practical immortality.

  12. #12
    Banned Beazy's Avatar
    10+ Year Old Account
    Join Date
    Dec 2010
    Location
    Dallas, TX
    Posts
    8,459
    Just this weekend I watched a video of a group of scientists who created a bunch of small drones that could fly around in patterns. They would take a stick and poke on of the drones flying by and fuck the whole thing up, but some how, all the drones were able to think together and correct the flight pattern back to normal. It was insane.

    That has to be some amazing level of AI. I will look in my youtube history vids and post a link when I find it.

  13. #13
    there are reckoned to be about 1080 atoms in the observable universe, for instance
    I'm not convinced

  14. #14
    Quote Originally Posted by Attackrabbit View Post
    I'm not convinced
    It's actually 1080. It's impressive how much difference that makes, no? lol.

  15. #15
    Deleted
    Good, the faster this -and other tech- advances the sooner we'll be forced to implement real fundamental change. Otherwise we'll just play status quo under different labels and guises.

  16. #16
    Immortal Shadochi's Avatar
    7+ Year Old Account
    Join Date
    Aug 2015
    Location
    The Shit Throne
    Posts
    7,802
    Quote Originally Posted by Gadion View Post
    I'd be more interested in real life applications that have less defined rule sets. Once an AI that knows nothing gets to learn anything by itself without guidance, that would be the birth of the AI species' dominance on Earth
    You mean just like humans can learn anything by themselves without any guidance? Oh wait, they can't...
    #1 Hype-Thread Shitposter - Overlord of the Hypethread

  17. #17
    Quote Originally Posted by Attackrabbit View Post
    I'm not convinced
    Quote Originally Posted by Attackrabbit View Post
    I'm not convinced
    I'm sure he meant 10 or 2 to the 1080th power.

    This is a quite interesting article, and what is happening with GO mimics what is happening in the Chess world with computers. Chess computers crush the best GMs. And if you put them on "do not use opening library of moves", then... well they basically recreate and refine the library. This is what happened in a recent Chess Computer tournament.

    I've never seen an extensive description of how the computers actually came up with such good moves. Thank you for this!

  18. #18
    The Lightbringer Cerilis's Avatar
    10+ Year Old Account
    Join Date
    Mar 2011
    Location
    Germany
    Posts
    3,191
    I, for one, welcome our new robot overlords.

  19. #19
    Deleted
    Quote Originally Posted by PrimaryColor View Post
    Libratus from earlier this year already showed AI can beat poker pros. Even bluffing has a rhyme and rhythm that can be exploited.
    Optimal Bluffing is actually very procedural and mathematical despite its depiction in popular culture as being "human". There's an optimal randomized percentage for every combination of cards-the only barrier to perfect play are the very large number of permutations making correct strategy difficult to calculate for humans. This is what computers have always been good at.

    I'm not sure AI is actually making as much progress as people seem to think. Computers are really good at working through permutations quickly because they don't have to worry about leaving the iron on. They don't seem to be making that much progress towards replicating human learning: see youtube's idiot AI banning random and arbitary videos.

  20. #20
    Dreadlord Gadion's Avatar
    10+ Year Old Account
    Join Date
    Jun 2013
    Location
    I Live On The Web
    Posts
    842
    Quote Originally Posted by Shadochi View Post
    You mean just like humans can learn anything by themselves without any guidance? Oh wait, they can't...
    Err, yes they can. Someone that taught themselves something is called an autodidact. I am not trying to say AI isn't impressive, I'm just saying that it'll get *really* impressive when they start to self-direct.

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •