[Boards: 3 / a / aco / adv / an / asp / b / bant / biz / c / can / cgl / ck / cm / co / cock / d / diy / e / fa / fap / fit / fitlit / g / gd / gif / h / hc / his / hm / hr / i / ic / int / jp / k / lgbt / lit / m / mlp / mlpol / mo / mtv / mu / n / news / o / out / outsoc / p / po / pol / qa / qst / r / r9k / s / s4s / sci / soc / sp / spa / t / tg / toy / trash / trv / tv / u / v / vg / vint / vip / vp / vr / w / wg / wsg / wsr / x / y ] [Search | Free Show | Home]

AlphaGo goes 60-0 vs top GO players including Ke jie

This is a blue board which means that it's for everybody (Safe For Work content only). If you see any adult content, please report it.

Thread replies: 36
Thread images: 6

https://www.youtube.com/watch?v=b_24iaUMRFs

AlphaGo is no the greatest GO playing entity on earth. Having gone 4-1 vs Lee Sedol, it is now 60-0 against the best players in the world.
>>
File: delet this.png (278KB, 740x732px) Image search: [Google]
delet this.png
278KB, 740x732px
>>8591444
COMPUTERS MUST NOT BE SMARTER THAN HUMANS
>>
>>8591445
>""""""""""""""""""""SMARTER""""""""""""""""""""

Reminder that computers are neither smart nor intelligent.
>>
>>8591444
Is this like Deep Blue VS Kasparov so is there Google techs behind the scenes adjusting AlphaGo on the fly and consistently changing the rules to narrow the scope of the 'win''
>>
>>8591445
What I consider "smart" really has nothing to do with being good at very specific tasks. We already know that computers are better at very specific tasks than humans, that is by no means a surprise.

Actual intelligence is figuring shit out on your own, without having someone program you. Even though DNNs are usually marketed as magic, they are not magic. Getting one to work requires a lot of work still, you need to carefully design the layout. The next step from DNNs would be self-organizing DNNs that figure out what kind of layout is best suited for a task. I'm not sure what the state there is, but I guess not very far. It should limit performance severely.
>>
>>8591509
>>8591445
It is though. This success is a clear example of the overwhelming superiority of electric transistors as compared to human neurons. It is able to process an unimaginable number of games and communicate in a direct way via fast interfaces.

This is actually more impressive than being more intelligent or less intelligent. It is something even better, fundamentally superior in what seems to be every way.

A human brain with 250 IQ is not as impressive as what alphago can do in terms of processing speed and data transference. This should scare the fuck out of everyone
>>
>>8591574
>This should scare the fuck out of everyone
It only scares the fuck out of people who don't know how computers work.
That's like saying a hammer should scare the fuck out of everyone because no human being can bang a nail as hard as it can. It's just a tool.
>>
>>8591444
Oh yeah, can it explain the meaning of life or how a vagina feels like?

Checkmate atheists
>>
go weeaboos btfo

their game has been reduced to tic tac toe.


same with chessfags, fuck them too
>>
>>8591444
If an infinite number of monkeys were to slam their fists randomly on controllers playing GO against a computer which makes the MOST OPTIMAL move everytime, is it not a certainty that one of the monkey's would win eventually? Now extend that to best of 3. Eventually a monkey would randomly beat a computer at least 2/3 times, and therefore win the series.

Therefore the game is nothing more than a game of chance. You may as well roll dice and call it a day.
>>
>>8591574

This computer doesn't understand shit at all, it's not intelligent.
>>
File: hissmileandchunnihairGOne.jpg (723KB, 1280x1390px) Image search: [Google]
hissmileandchunnihairGOne.jpg
723KB, 1280x1390px
>>8591444
>4-1
How did AlphaGo lose? Out of pity for puny humans?
>>
>>8591574
Computers are still shit at pattern matching. Humans are still king when it comes to this.

What we need to do is figure out how to interface transistors with the human brain.
>>
>>8591885
Okay buddy, lets see you out-perform grep.

I'm waiting.
>>
>>8591826
All that stress from the match made his hair shrink. EITHER THAT OR AI SHRINKS HAIR
>>
why do people act like this means that robots are better than humans at go?

all this means is that robots have become better than humans on small go boards.

Go can be played on many different sizes of go board.

Humans have a genuine understanding of Go and so would be able to use their heuristics to play on a 100x100 board if they really wanted to and they'd be somewhat worse than on 19x19 or smaller but alphago would perform dreadfully.

So really, this isn't a much bigger deal than deep blue
>>
>>8592180
>but alphago would perform dreadfully
citation needed
>>
>>8591574
see look at this idiot.

"omg the power of silicon defeaing neurons! the computer now literally understands the game better than humans!"

fuck off

>>8591826
read this blog
https://gogameguru.com/lee-sedol-defeats-alphago-masterful-comeback-game-4/

but if you don't know the rules of Go then you still won't really understand why lee won
>>
>>8591885
>Computers are still shit at pattern matching. Humans are still king when it comes to this
not really, judging by the sheer number of retards that still walk the earth believing in absolute nonsense
people don't readily recognize patterns unless they are taught to recognize them
>>
>>8592185
lphaGo has been trained on a standard full-size Go-board. Both it's move selection and position evaluation convolutional neural networks have been trained extensively on the basis of 130,000 amateur games (19x19) and millions of self-play games. Both neural networks depend heavily on the exact board position, especially the position evaluation. Both neural networks have been trained by reinforcement learning using the outcome of the game on a standard 19x19 board.

furthermore , part of alpha go still uses monte carlo tree search , which is totally dependent on board size.
>>
>>8592193
you only need a few intelligent people for the species as a whole to advance.
>>
>>8591445
Quick! How much is 466832 times 756432?
If this took you more than 0.01 seconds you must destroy your calculator at once.
>>
File: b834fd16e3d51d43557a9df1a31887cd.jpg (98KB, 1920x1080px) Image search: [Google]
b834fd16e3d51d43557a9df1a31887cd.jpg
98KB, 1920x1080px
You're in the lab and this guy laughs at your research paper. What do you do?
>>
>>8591815
humans dont understand shit either

We just reproduce and survive
>>
where do I go to learn Go strategies?

I get the game but I basically just make ladders with the CPU and it comes up with ways to eat my pieces
>>
>>8591574
it scares people who don't understand what they're talking about, like you
>This success is a clear example of the overwhelming superiority of electric transistors as compared to human neurons.
complete and utter nonsense
>>
>>8594294
https://online-go.com/puzzles

Go to the Puzzles tab. There's a lot of tutorial packs that are pretty good.
>>
File: DQeCW.gif (2MB, 380x214px) Image search: [Google]
DQeCW.gif
2MB, 380x214px
They're doing Starcraft 2 next which should be really cool.
>>
>>8594319
Thank you
>>
>>8594326
>the most highly optimized zerg rush ever performed
>>
File: 1419660590181.jpg (258KB, 800x756px) Image search: [Google]
1419660590181.jpg
258KB, 800x756px
>>8592302
>>
>>8594326
>switching from Go to starcraft
does google just really hate Koreans or something
>>
>>8592918
some of us do not reproduce
>>
>>8591579
>no human being can bang a nail as hard as it can
There is a "your mother" joke somewhere in there, but I can't find it (AlphaGo could, though).
>>
>>8592195
It's *exactly* the same case for humans. Go is a game of memory and pattern recognition. No one really "understands" the game; strong players just acquire a database of millions of positions and learn their (sometimes minute) strategic differences. This is what we celebrate as "human intuition". Actual reading ahead is helped by this database, but still not fundamentally different from MC.
>>
>>8591885
>computers are shit at pattern matching
Depends on the pattern and how the data is represented.
Thread posts: 36
Thread images: 6


[Boards: 3 / a / aco / adv / an / asp / b / bant / biz / c / can / cgl / ck / cm / co / cock / d / diy / e / fa / fap / fit / fitlit / g / gd / gif / h / hc / his / hm / hr / i / ic / int / jp / k / lgbt / lit / m / mlp / mlpol / mo / mtv / mu / n / news / o / out / outsoc / p / po / pol / qa / qst / r / r9k / s / s4s / sci / soc / sp / spa / t / tg / toy / trash / trv / tv / u / v / vg / vint / vip / vp / vr / w / wg / wsg / wsr / x / y] [Search | Top | Home]

I'm aware that Imgur.com will stop allowing adult images since 15th of May. I'm taking actions to backup as much data as possible.
Read more on this topic here - https://archived.moe/talk/thread/1694/


If you need a post removed click on it's [Report] button and follow the instruction.
DMCA Content Takedown via dmca.com
All images are hosted on imgur.com.
If you like this website please support us by donating with Bitcoins at 16mKtbZiwW52BLkibtCr8jUg2KVUMTxVQ5
All trademarks and copyrights on this page are owned by their respective parties.
Images uploaded are the responsibility of the Poster. Comments are owned by the Poster.
This is a 4chan archive - all of the content originated from that site.
This means that RandomArchive shows their content, archived.
If you need information for a Poster - contact them.