Go master Lee Se-dol says he quits, unable to win over AI Go players

Post by selestify » Fri, Nov 29, 2019, 5:44 AM UTC

dfan wrote: post #522

breatheoften wrote: post #470

Earlier quoted context omitted.

This is interesting to me ... how exactly is this assessed ...? Do people keep around versions of “alpha go year 2017” and play against it in order to measure human improvement over time? If the basis for observing improvement has become “I can beat old versions of the ai more reliably than I used to be able to” or if I have learned to beat players who have not studied alpha zero I suppose that’s a form of usefully l…

I admit that I do not have a quantitative measure to support my claim (as you note, constructing one is difficult). But qualitatively: 1) People have learned a lot from new engines: joseki (corner patterns), general strategy (e.g., moves on the side are now considered less valuable, making large moyos (largely empty space loosely surrounded by your stones) is less attractive because AIs have demonstrated that they're…

Post by palerdot » Fri, Nov 29, 2019, 4:43 PM UTC

jackcviers3 wrote: post #536

braythwayt wrote: post #527

Earlier quoted context omitted.

> Imagine a world where no human contribution matters because a different species already thought of it You are begging the question in in the original sense of the phrase "begging the question." You start by stating your conclusion as an axiom! You state that "no human contribution matters because a different species already thought of it" as if that is necessarily true. It isn't. We choose what matters. I play Bach…

Your playing doesn't "matter" in the way Bach's original composition "matters." One of these things has been remembered for hundreds of years, and one of these things would never have been remembered unless you reinterpreted the work in some novel way that reached millions of people. The fact that Bach's compositions can be encoded as a number doesn't mean that it wouldn't take a genius of Bach's level to deserialize…

Post by nelgaard » Sat, Nov 30, 2019, 2:00 AM UTC

jv22222 wrote: post #425

breatheoften wrote: post #374

I sympathize. It used to be that you might be able to believe that there is some kind of art behind go, some sort of abstract beauty to it, and that the pursuit of this beauty is the path to being good at go ... But the defeat of the tactics born from this mindset by MCTS at least for now lay bare the fact that the path to being good at go is actually to probabilistically sample the phase space of the game and perfor…

I’m now curios, can a human design a game that a computer can not beat a human at.

Post by osyed1 » Sat, Nov 30, 2019, 8:39 AM UTC

arunix wrote: post #511

andrepd wrote: post #439

Earlier quoted context omitted.

https://en.wikipedia.org/wiki/Arimaa Also, relevant xkcd: https://xkcd.com/1002/

This hasn't been true since 2015, when the software became decisively stronger than human players. https://www.kingpinchess.net/2015/07/arimaa-game-over/

Post by breatheoften » Sat, Nov 30, 2019, 10:59 PM UTC

pyrale wrote: post #509

breatheoften wrote: post #374

I sympathize. It used to be that you might be able to believe that there is some kind of art behind go, some sort of abstract beauty to it, and that the pursuit of this beauty is the path to being good at go ... But the defeat of the tactics born from this mindset by MCTS at least for now lay bare the fact that the path to being good at go is actually to probabilistically sample the phase space of the game and perfor…

I don't want to break your ideal, but the description of go you make died decades ago. For a long time now, top go has been about fighting in the chuban, and there is no grand theory about that, only grinding your mind to read more combinations than your opponent. Another correction about what you said is that computer go has brought tremendous evolutions to the pro scene. Josekis that were accepted for centuries hav…