AlphaGo Defeats World's Best Human Player

AI Go Program Performs Better Than Expected

Anything that doesn't fit anywhere else below.

Moderators: Calilasseia, ADParker

AlphaGo Defeats World's Best Human Player

#1  Postby Calilasseia » Mar 09, 2016 2:18 pm

New Scientist has just notified me of this report, on the contest between the AlphaGo artificial intelligence Go playing program, and Lee Sedol, currently considered the world's best human player.

Apparently, AlphaGo has won the first of the five scheduled games in the challenge match, and Lee Sedol is reported to have said “I didn’t think AlphaGo would play the game in such a perfect manner.” after the defeat.

The result is all the more striking, because even a number of experts in the AI field, regarded this level of proficiency on the part of an AI Go program to be a decade away.

It's possible that Lee could recover and take the next four games. But, AlphaGo has already defeated a previous human Go player of stature, European champion Fan Hui, a straight 5-0. If Lee Sedol is going to beat AlphaGo, it appears he has his work cut out for him.
Signature temporarily on hold until I can find a reliable image host ...
User avatar
Calilasseia
RS Donator
THREAD STARTER
 
Posts: 21697
Age: 56
Male

Country: England
United Kingdom (uk)
Print view this post

Ads by Google




Re: AlphaGo Defeats World's Best Human Player

#4  Postby Calilasseia » Mar 10, 2016 7:06 pm

UPDATE: AlphaGo is now leading 2-0.
Signature temporarily on hold until I can find a reliable image host ...
User avatar
Calilasseia
RS Donator
THREAD STARTER
 
Posts: 21697
Age: 56
Male

Country: England
United Kingdom (uk)
Print view this post

Re: AlphaGo Defeats World's Best Human Player

#5  Postby tuco » Mar 10, 2016 8:26 pm

12,13,15, next matches.

At yesterday’s post-game press conference, Lee looked shell-shocked. Today he seemed resigned. One reporter asked what AlphaGo’s weaknesses are? Lee laughed: “Well, I guess I lost the game because I wasn’t able to find out.”


Sound like when Kasparov was against Deep Blue in 1997.

One day it will be able to adapt, counter exploits, and everyone will be fucked ;)
---

Update: Why this week’s man-versus-machine Go match doesn’t matter (and what does)

The win sent shock waves through the Go community. “When I saw the games, I was very surprised because the computer was playing like a human,” says Hajin Lee, secretary general of the International Go Federation in Tokyo. “If you had not told me, I could not tell from the moves which one was the computer.”

Then came a process of acceptance. “To me it was a relief,” says Frank Lantz, the director of the New York University Game Center in New York City and an enthusiastic Go amateur. “Go had become this mythic bulwark against artificial intelligence. People got used to this proverb: ‘Yes, computers can play chess but they can’t play Go.’ I don’t think it was ever true. We can finally put away the last self-delusion that there is some magical quality to Go that makes it intrinsically human.”


In games, human exceptionalism may hold out a while longer in poker. Computers are already close to playing mathematically optimal strategies in some versions of two-handed poker. But good human poker players possess an extra skill: the ability to read opponents’ weaknesses—their deviations from an optimal strategy—and exploit them. Computers can’t do that yet, but in principle it’s just the sort of thing a deep neural net ought to be able to master, says Nikolai Yakovenko of Twitter New York in New York City, a former poker pro who is working on artificial-intelligence poker software. Then another bastion will fall.


http://www.sciencemag.org/news/2016/03/ ... -what-does
tuco
 
Posts: 13802

Print view this post

Re: AlphaGo Defeats World's Best Human Player

#6  Postby Calilasseia » Mar 13, 2016 4:27 am

AlphaGo has now won the tournament 3-0.

The remaining two matches will still be played, simply to see if AlphaGo can achieve a 5-0 result.
Signature temporarily on hold until I can find a reliable image host ...
User avatar
Calilasseia
RS Donator
THREAD STARTER
 
Posts: 21697
Age: 56
Male

Country: England
United Kingdom (uk)
Print view this post

Re: AlphaGo Defeats World's Best Human Player

#7  Postby tuco » Mar 13, 2016 8:54 am

From the article:

Does this mean AI is now smarter than us and will kill us mere humans? It certainly does. Mere humans can't evaluate complex trade arrangements like TTIP or military intelligence concerning WMD for example, said researchers at British computer company DeepMine. AlphaGo votes Mr Trump.


;)
tuco
 
Posts: 13802

Print view this post

Ads by Google



Re: AlphaGo Defeats World's Best Human Player

#9  Postby tuco » Mar 13, 2016 11:20 am

When I say 'thought' and 'realisation' I just mean the output of #AlphaGo value net. It was around 70% at move 79 and then dived on move 87


Lee Sedol wins game 4!!! Congratulations! He was too good for us today and pressured #AlphaGo into a mistake that it couldn’t recover from


https://twitter.com/demishassabis

Grats Mr Sedol. Pressuring AI to mistakes :)
tuco
 
Posts: 13802

Print view this post

Re: AlphaGo Defeats World's Best Human Player

#10  Postby I'm With Stupid » Mar 14, 2016 6:59 am

Impressive machine, but could it win a game of Mouse Trap?
Image
User avatar
I'm With Stupid
 
Posts: 8992
Age: 34
Male

Country: UK
Jolly Roger (arr)
Print view this post

Re: AlphaGo Defeats World's Best Human Player

#11  Postby tuco » Mar 16, 2016 4:43 pm

4:1

#AlphaGo wins game 5! One of the most incredible games ever. To comeback from the initial big mistake against Lee Sedol was mind-blowing!!!


https://twitter.com/demishassabis

--

Moving on: DeepMind Challenges for StarCraft

StarCraft is a great testbed for AI, because it presents many of the challenges necessary for performing real-world tasks. As part of my dissertation, I classified StarCraft in terms of Russell and Norvig’s task environment properties. The results from this analysis are shown in the figure below, with real-world properties highlighted as bold. The only difference between StarCraft and real-world activities such as taxi driving is that StarCraft is a deterministic environment. This example lists Chess as a conventional AI challenge, but all of the properties are the same for Chess and Go.


http://gamasutra.com/blogs/BenWeber/201 ... rCraft.php
tuco
 
Posts: 13802

Print view this post

Re: AlphaGo Defeats World's Best Human Player

#12  Postby JoeB » Mar 16, 2016 11:13 pm

How long before such systems will be used in combat scenarios? I can imagine a similar AI outsmarting every human general in combat, no?
User avatar
JoeB
RS Donator
 
Name: Johan
Posts: 1463
Age: 31
Male

Country: European Union (NL)
European Union (eur)
Print view this post

Re: AlphaGo Defeats World's Best Human Player

#13  Postby CdesignProponentsist » Mar 17, 2016 12:10 am

JoeB wrote:How long before such systems will be used in combat scenarios? I can imagine a similar AI outsmarting every human general in combat, no?


Combat is orders of magnitude more complex than Go and comes with hugely unpredictable outcomes as well as trying to remain unpredictable. I don't think AI will be involved in more than an aid for some time, but I wouldn't rule out it eventually out performing human intelligence.
"Things don't need to be true, as long as they are believed" - Alexander Nix, CEO Cambridge Analytica
User avatar
CdesignProponentsist
 
Posts: 12146
Age: 51
Male

Country: California
United States (us)
Print view this post


Re: AlphaGo Defeats World's Best Human Player

#15  Postby Calilasseia » Mar 17, 2016 12:46 am

AlphaGo has now won the series 4-1.
Signature temporarily on hold until I can find a reliable image host ...
User avatar
Calilasseia
RS Donator
THREAD STARTER
 
Posts: 21697
Age: 56
Male

Country: England
United Kingdom (uk)
Print view this post

Ads by Google


Re: AlphaGo Defeats World's Best Human Player

#16  Postby tuco » Mar 17, 2016 8:28 am

Following a link from Mr Hassabis tweets.

Demis Hassabis, master of the new machine age

The victories have a human mastermind in Demis Hassabis, co-founder and chief executive of DeepMind. He describes Mr Lee as the “Roger Federer of Go”, and for some the computer program’s achievement is akin to a robot taking to the lawns of Wimbledon and beating the legendary tennis champion.


“One of the curiosities of the phenomenal progress we’re making with AI is that it looks as though we have a world champion at Go, but we don’t have a computer that can physically move the Go pieces,” he says. Mr Federer will not face a similar challenge just yet.


http://www.ft.com/cms/s/0/630bcb34-e6b9 ... z42QXoGHS5
----

I see I made mistake, though initially I wrote Mr Lee for some reason then edited it, calling Mr Lee Sedol Mr Sedol. Lee Sedol .. how does that work?

https://en.wikipedia.org/wiki/Korean_name

Oh they have it backwards, or we do.

In English publications, usually Korean names are written in the original order, with the family name first and the given name last. This is the case in Western newspapers. Koreans living and working in Western countries have their names in the Western order, with the given name first and the family name last. The usual presentation of Korean names in English is similar to those of Chinese names and differs from those of Japanese names, where they, in English publications, are usually written in a reversed order with the family name last.


How embarrassing. Apologies Mr Lee.
tuco
 
Posts: 13802

Print view this post

Re: AlphaGo Defeats World's Best Human Player

#17  Postby Hardcoreathiest » Mar 21, 2016 3:21 am

CdesignProponentsist wrote:
JoeB wrote:How long before such systems will be used in combat scenarios? I can imagine a similar AI outsmarting every human general in combat, no?


Combat is orders of magnitude more complex than Go and comes with hugely unpredictable outcomes as well as trying to remain unpredictable. I don't think AI will be involved in more than an aid for some time, but I wouldn't rule out it eventually out performing human intelligence.



Sooner than that. AlphaGo is an applied test for a general deeplearning - re-enforcement system that can outperform humans on 3d games. The next few modules they are adding will add some of the remaining key brain features and also allow it to transfer its learning from one scenario to another. And given that its all standard software, anything one system learns under training can be transferred to a central database. This means exponential increases in experience across hundreds of training scenarios could occur in weeks. Then when the system is released to the public for their own use there will be exponential increases in training data. This integration of skills is something that currently takes human societies of millions of people generations to pull off. These systems will completely transform the world very quickly
Hardcoreathiest
 
Name: Joe Alan
Posts: 60

Country: ireland
Ireland (ie)
Print view this post

Re: AlphaGo Defeats World's Best Human Player

#18  Postby I'm With Stupid » Mar 21, 2016 3:33 am

Hmm, an AI machine that would think nothing of sacrificing vast number of people to win the "game" in charge of military tactics? Fun.
Image
User avatar
I'm With Stupid
 
Posts: 8992
Age: 34
Male

Country: UK
Jolly Roger (arr)
Print view this post

Re: AlphaGo Defeats World's Best Human Player

#19  Postby tuco » Mar 21, 2016 4:00 am

In defence of AlphaGo:

----
It becomes self-aware at 2:14 a.m. Eastern time, August 29th. In a panic, they try to pull the plug.
Sarah Connor: Skynet fights back.


and it goes downhill from here:

Yes. It launches its missiles against the targets in Russia.
John Connor: Why attack Russia? Aren't they our friends now?
The Terminator: Because Skynet knows that the Russian counterattack will eliminate its enemies over here.


That is clever of Skynet. Output of value net around 99% ;)
---

who cant defend itself, yet.
tuco
 
Posts: 13802

Print view this post

Re: AlphaGo Defeats World's Best Human Player

#20  Postby Hardcoreathiest » Mar 23, 2016 3:47 pm

ridiculous scenarios. Any Ai system that develops recurrently to include greater memory over events will be even more consistent with itself than a direct programmed system. i.e. If it has human protection values programmed from the start it is extremely unlikely to deviate later on.
Hardcoreathiest
 
Name: Joe Alan
Posts: 60

Country: ireland
Ireland (ie)
Print view this post

Next

Return to General Science & Technology

Who is online

Users viewing this topic: No registered users and 1 guest