Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

AlphaGo beats Lee Sedol 3-0 [video]

youtube.com

Post reply on HN
Print view View original

281–290 of 428 posts

Re: AlphaGo beats Lee Sedol 3-0 [video]

#281

Post by Anderkent »

faitswulff wrote: post #215
samizdatum wrote: post #136

Earlier quoted context omitted.

You're implying that if something can be outdone it doesn't have a purpose, which seems to rule out purposes for pretty much everything. I'm sure there's always someone that can write books or maths papers or symphonies better than you. I don't think this robs you of purpose, unless your purpose is to be the absolute best at something. Anyway, I find it curious that you would say logic is a quintessentially human tra…

The difference between being outdone by another human and being outdone by a computer is that the computer's efforts are nearly infinitely reproducible, given the processing power. So a more apt analogy would be if there was someone inside every cellphone who could write books, papers, or symphonies better than you. That day is coming.

And it would be great. Think of all the great symphonies and books!

Re: AlphaGo beats Lee Sedol 3-0 [video]

#282

Post by hasenj »

tome wrote: post #210
mathgenius wrote: post #164

Earlier quoted context omitted.

It seems like in all three games, AlphaGo plays a thick move that looks a bit baffling to the commentators, but then miraculously those moves become hugely profitable somewhat later on in the game.

Is it possible that all of AlphaGo's strength is in these baffling moves? If humans played enough games against AlphaGo and discovered how to counter the baffling moves, is it plausible that AlphaGo's strength would be lost?

It seems to me that AlphaGo's main strengths are:

- It doesn't make mistakes due to pressure or fatigue; it's always playing at its sharpest

- It can read deeper than humans can

- It seems to have better judgement than most current professional players

Re: AlphaGo beats Lee Sedol 3-0 [video]

#283

Post by mratzloff »

astrofinch wrote: post #221
landryraccoon wrote: post #55

Earlier quoted context omitted.

Well, it would have to have a motivation to do so. Evolution has put complex motivations into human beings for billions of years, self preservation being chief among them. Even if we put motivation for self preservation into an AI, we might not do as well as nature did, leaving the AI open to self destruction or shutdown by humans - simply because the AI has no motivation not to allow humans to turn it off. Human des…

Self-preservation falls out of almost any other goal you give an AGI. If I program my AGI with the goal of making my startup succeed, and the AGI thinks it can help, then me shutting it off is a potential threat to my startup's success. So of course it will try to prevent that the same way it would try to prevent any other threat to my startup's success. World domination is a similar situation. For any goal you give…

It has to be aware that it can be shut down and have the capacity to prevent that. AlphaGo doesn't know it can be shut down and therefore couldn't "care" less--even if it was shut down in the middle of a game.

Re: AlphaGo beats Lee Sedol 3-0 [video]

#284

Post by JibberMeTimbers »

21 wrote: post #95
seanwilson wrote: post #71

Earlier quoted context omitted.

I read that too. It's interesting StarCraft is suggested instead of any classic board games. It sounds like Go is the pinnacle of board games then. Would StarCraft be the last game AI has to beat?

No, after StarCraft they suggested 3D games (eg: Call of Duty, ...). Mind you, this is playing based on the pixels, not based on direct access to the game structures as bots do today. The from playing Call of Duty to controling a robot with a gun in the real world it's a small step :)

I don't like the idea of FPS'. We'd expect them to behave the humans when they play FPS - they run looking forward and will occasionally look left or right and follow some predetermined pathing that they are familiar with.

With an AI, if we base it on pixels, what they will do is spin around 360 degrees about 60 times a second while moving forward so that it can maximize all of the pixels inputted at a time. It would compare it against the stored level design and shoot at anything that is off (or against character models). I can just see that it's not going to behave like anything we'd expect.

Re: AlphaGo beats Lee Sedol 3-0 [video]

#285

Post by daveguy »

Madmallard wrote: post #249
m0g wrote: post #198

Earlier quoted context omitted.

Your pocket calculator is superhuman at arithmetic. Alphago is superhuman at go. It's not a big deal.

The way in which it is superhuman matters. The calculator uses simple mechanical algorithms. Alpha Go uses a completely novel approach to deep learning that can likely be applied to many other systems/problems.

As long as those systems/problems include a grid based problem space where the goal is to successfully place stones restricted by a limited set of rules.

Ok flippancy aside, there are two problems that make techniques like this single-domain: network design and network training.

The design, uses multiple networks for different goals: board eval (what boards look good) and policy (which moves to focus on). Those two goals, eval and policy, are very specific to go. Just like category layers are specific to vision and LSTM is specific to sequence learning.

Network training is obviously hugely resource intensive -- and each significantly complex problem would need such intensity.

It is amazing the variety of problems DNNs have been able to do well in. However, the problem of network design and efficient training are significant barriers to generalization.

When network design can be addressed algorithmically I think we may have an AGI. However, that is a significant problem where you automatically add another layer of computational complexity so it is not on the immediate horizon and may be 50+ years down the road.

Re: AlphaGo beats Lee Sedol 3-0 [video]

#286

Post by supercheetah »

camperman wrote: post #250
tzaman wrote: post #193

Earlier quoted context omitted.

Watch our for nvidia drivers :)

And be careful of systemd. Most days you'll wake up in 2 seconds but occasionally you'll be comatose for the whole day with no way to find out what's wrong.

That's alcohold that you're thinking of.

Re: AlphaGo beats Lee Sedol 3-0 [video]

#287

Post by webreac »

oneeyedpigeon wrote: post #133
pmontra wrote: post #88

Earlier quoted context omitted.

I agree. Lee lost a title 3-2 to Ke Je recently but he lost very convincingly against AlphaGo. Ke Je is not that stronger than Lee. He's going to lose too but he's 18 so I understand that he's feeling he can do anything. My feeling is that AlphaGo could be a couple of stones stronger than the best pros. Considering that dans are capped at 9 but there are 9 dans that always win against some other 9 dans, the best pros…

Do you know what the current limit is on a handicap that a professional player can overcome? Is there a known theoretical limit? That's a possible 'endgame' for go ai.

I have read that a go world champion playing against god would need between 4 and 5 stones of handicap.

Re: AlphaGo beats Lee Sedol 3-0 [video]

#288

Post by megalodon »

YeGoblynQueenne wrote: post #275
megalodon wrote: post #261

Earlier quoted context omitted.

The same statement holds for humans too. I don't think you can teach a person how to play Go and expect him/her to learn anything else than how to play Go.

I'm not talking about learning the rules of the game. You don't need an AI to model the rules of the game. You need an AI to model the winning strategy. Winning strategy is what humans generalise to other domains and computers don't. There's a lot that suggests that humans and machine learning algorithms learn in very different ways. For instance, by the time a human can master a game like Go they can also perform im…

I was just contradicting your statement.

A human brain (or any other animal brain for that matter) is almost infinitely more advanced and computationally efficient than state of the art machine intelligence, even without taking things like thoughts, emotions and dreams - which we currently do not understand at all - into account.

It's a huge accomplishment for machines to be able to win over humans in games like chess and go and , but these are games originally designed for humans - by humans - to be played recreationally and I think we shouldn't read too much into it.

Re: AlphaGo beats Lee Sedol 3-0 [video]

#289

Post by iopq »

agildehaus wrote: post #115
theroof wrote: post #73

Is anyone also asking themselves when they'll be able to play against this level of AI on their mobile phone? Or formulated differently: when will an "AlphaGo" (or equivalent) app appear in the play/app store? In 2 years? In 1 year? In 3 months?

A version of AlphaGo that beat Fan Hui ran on a single machine, but a very beefy one (48 CPUs and 8 GPUs). Unless they can optimize it quite a bit, we won't be seeing this on a mobile phone anytime soon. Perhaps you'll be able to pay a good hourly rate for the cloud version.

A mobile phone version will still trounce the current AIs like CrazyStone considering the single machine version is competitive with the cluster (can take some games off of it)

Re: AlphaGo beats Lee Sedol 3-0 [video]

#290

Post by mtgx »

empath75 wrote: post #225
YeGoblynQueenne wrote: post #205

Earlier quoted context omitted.

>> What would you consider to be "true AI"? At what point are you willing to say, "Okay, that's it, computers are just plain smarter than we are?" Look, there's no doubt that computers can outperform humans in specific tasks. There's no doubt that AlphaGo is intelligent when it comes to Go, but on the other hand it would be completely incapable of tackling a different congitive task- say, language, or vision, or disc…

Actually alpha go can be trained for vision or any other task. It's not specialized to win go.

I think AlphaGo is specialized to play Go. You must be thinking DeepMind, on which AlphaGo is based.
Post reply on HN