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

AlphaGo beats the world champion Lee Sedol in first of five matches

twitter.com

Post reply on HN
Print view View original

541–550 of 596 posts

Re: AlphaGo beats the world champion Lee Sedol in first of five matches

#541

Post by ogrisel »

argonaut wrote: post #500
ogrisel wrote: post #472

Earlier quoted context omitted.

...and it's possible to train learning agents to sense and interact with a world described by high dimensional continuous vector spaces, for instance using conv nets (for sensing audio / video signals) and actor-critic to learn an continuous policy: http://arxiv.org/abs/1509.02971 The fact that the (reinforcement) learning problem is hard or not is not directly related to whether the observation and action spaces are…

There is a near infinite number of such spaces.

I don't understand why would the "number of spaces" matters. What matters is can you design a learning algorithm that performs well in interesting spaces such as:

- discrete spaces such as atari games and go, - continuous spaces such as driving a car, controlling a robot or bid on a ad exchange.

Re: AlphaGo beats the world champion Lee Sedol in first of five matches

#542

Post by koder2016 »

Eliezer wrote: post #30
EGreg wrote: post #17

Does this mean in the next few decades, computers will make better sex partners and companions than any human?

Worrying about the effect of strong AI on sexual relationships is like worrying about the effect on US-Chinese trade patterns if the Moon crashes into the Earth.

...or like worrying about being on board of a heavier than air aircraft (surely that's impossible).

Re: AlphaGo beats the world champion Lee Sedol in first of five matches

#543

Post by EliRivers »

EliRivers wrote: post #155
atemerev wrote: post #77

Earlier quoted context omitted.

The only remaining are language-related. Natural languages are the next focal point of AI research.

I see "AI" doing well in games that have simple inputs, a limited range of legal outputs, and relatively easy "goodness" measures. I see nothing that might be able to tell us why gravitational mass is the same as inertial mass, for example, or any moves in that direction. This "AI" is good at simple games.

As an addendum to my comment, a number of people working at Deepmind agree with me.

Re: AlphaGo beats the world champion Lee Sedol in first of five matches

#544

Post by simonh »

Houshalter wrote: post #448
fma wrote: post #264

Earlier quoted context omitted.

As far as I know the goal post of Turing test has never moved.

That's because it hasn't been beaten yet! As soon as a chatbot beats a turing test, there will be a lot of AI deniers come out and say that the Turing test doesn't measure ' real ' intelligence, or isn't a valid test for . I know this is true, because there are already a lot of people that think the Turing test isn't valid. They believe it could be beaten by a stupid chatbot, or deception on the part of AI. Just sear…

Viewed that way, I'll accept that beating Go represent progress. That's not the same as saying that it represents imminent evidence that singularity style strong AI is almost upon us, as suggested in the post I was replying to. In the long term it might turn out to represent very minimal progress towards that goal.

Re: AlphaGo beats the world champion Lee Sedol in first of five matches

#545

Post by nl »

areyousure wrote: post #450
baby wrote: post #441

Earlier quoted context omitted.

What kind of source do you want? It's a saying in the go community, people believed (including me) that a bot couldn't beat a human in our lifetime, some people had more extreme view and thought that it would never be possible.

Anything written. I'll be particularly happy with higher "quality" sources -- books, quotations in newspapers, etc. -- but honestly, I'm not that picky and will accept an anonymous comment on a random forum.

It seems unlikely that a computer will be programmed to drub a strong human player any time soon, Dr. Reiss said. ''But it's possible to make an interesting amount of progress, and the problem stays interesting,'' he said. ''I imagine it will be a juicy problem that people talk about for many decades to come.''[1]

Not quite what you are after, but it's pretty clear that he didn't think it would be beating the world champion in 14 years.

[1] NY Times, 2002, http://www.nytimes.com/2002/08/01/technology/in-an-ancient-g...

Re: AlphaGo beats the world champion Lee Sedol in first of five matches

#546

Post by argonaut »

nullc wrote: post #535
argonaut wrote: post #498

Earlier quoted context omitted.

They could. There was a different set of tricks that didn't work as well (greedy pretraining).

Lots of people tried and failed. Today lots of people-- ones with even less background and putting in less effort-- try and are successful. This is not a small change, even if it is the product of small changes.

Reconnecting to my original point way up-thread, my point is these "innovations" have not substantially expanded the types of models we are capable of expressing (they have certainly expanded the size of the types of models we're able to train), not nearly to the same degree as backprop/convnets/LSTMs did way back decades ago (this is important because AGI will require several expansions in the types of models we are capable of implementing).

Re: AlphaGo beats the world champion Lee Sedol in first of five matches

#547

Post by argonaut »

ogrisel wrote: post #541
argonaut wrote: post #500

Earlier quoted context omitted.

There is a near infinite number of such spaces.

I don't understand why would the "number of spaces" matters. What matters is can you design a learning algorithm that performs well in interesting spaces such as: - discrete spaces such as atari games and go, - continuous spaces such as driving a car, controlling a robot or bid on a ad exchange.

A really really large number of distinct decisions that need to be made. A car only needs to control a small set of actions (wheels, engine, a couple others I'm missing). A game player only needs to choose from a small set of actions (e.g. place piece at position X, move left/right/up/down/jump).

Re: AlphaGo beats the world champion Lee Sedol in first of five matches

#548

Post by argonaut »

nl wrote: post #536
fma wrote: post #270

Earlier quoted context omitted.

It's only difficult because no one threw money at it. It's like saying going to Mars is difficult. It is - but most of the technology is there already, just need money to improve what was used to go to the moon. If you asked people 10 years ago before the moon landing if it was possible, I too would agree it's impossible. But after that breakthrough it opened up the realm is possibilities. I see AlphaGo more of an in…

So are you arguing that superhuman-level performance in just a matter of engineering effort? Or am I missing something? I'm generally considered to be way over optimistic in my assessment of AI progress. But wow.. that's pretty optimistic!

I interpreted him as saying superhuman-performance at Go was just a matter of engineering effort, which I wholeheartedly agree with.

Re: AlphaGo beats the world champion Lee Sedol in first of five matches

#549

Post by awwducks »

forgotpwtomain wrote: post #320
jonah wrote: post #137

Earlier quoted context omitted.

It was live streamed and the archive is now up on The Official AGA Youtube Channel at: AlphaGo ?p vs Lee Sedol 9p, 0400 UTC (8pm PST)[1] [1] https://www.youtube.com/watch?v=6ZugVil2v4w

Thanks for the link. Myungwan Kim's commentary is superb, am I the only one thinking the american guy speaks a bit too much though?

I don't mind Andrew. He's a strong player in his own right so he has questions a strong amateur would have.

He inevitably asks questions you want Myungwan to answer.

Re: AlphaGo beats the world champion Lee Sedol in first of five matches

#550

Post by ogrisel »

argonaut wrote: post #547
ogrisel wrote: post #541

Earlier quoted context omitted.

I don't understand why would the "number of spaces" matters. What matters is can you design a learning algorithm that performs well in interesting spaces such as: - discrete spaces such as atari games and go, - continuous spaces such as driving a car, controlling a robot or bid on a ad exchange.

A really really large number of distinct decisions that need to be made. A car only needs to control a small set of actions (wheels, engine, a couple others I'm missing). A game player only needs to choose from a small set of actions (e.g. place piece at position X, move left/right/up/down/jump).

A human brain also has a limited number of muscles to control to interact with the world.
Post reply on HN