It is currently Tue Nov 19, 2019 11:24 pm

All times are UTC - 8 hours [ DST ]




Post new topic Reply to topic  [ 66 posts ]  Go to page Previous  1, 2, 3, 4  Next
Author Message
Offline
 Post subject: Re: Video series: Learning from AlphaGo
Post #21 Posted: Tue Jan 02, 2018 7:55 am 
Lives with ko

Posts: 143
Liked others: 45
Was liked: 33
Rank: EGF 3d
KGS: gennan
Tygem: gennan
OGS: gennan
Kaya handle: gennan
djhbrown wrote:
daal wrote:
I thought the explanations were clear and made sense
agree - so much so that they got me thinking and asking a question
At the moment, my understanding of gennan's dia 5 (reproduced below) is: black 3 could have two meanings (i didn't mean to suggest that black should only consider the top right after any white 4).

Black 1 and 3 ask white a question; white's answer at 4 also has two meanings - it either implies she thinks the outside is important, or it's a feint to busy black outside so white can take the inside... so black will reassess things and look for a move to keep her options open, making use of both black 1 and 3.
Attachment:
example4.php.png


I can only guess what AlphaGo and pros are thinking (from the continuations they play), but I see only one response by them after white 4 (=move 5 at E14). When only one move is possible for move 5, I see white 4 as a forcing move (or a rethorical question) rather than a question.

I kind of understood that you wanted to considering other options for black's move 5 to take the outside instead of the inside (so that move 4 can be considered a question instead of a forcing move), so I tried to find a reasonable alternative for move 5. But then from your later posts it seems that perhaps you weren't even asking me a question (or perhaps you were, but I got lost somewhere).

Perhaps you just wanted to confirm that move 4 is a forcing move with multiple meanings (instead of contemplating black's options for move 5)?
If that is the case, then I would say: yes, move 4 has multiple meanings. It's a kind of double attack. This kind of double attack is usually called "leaning" or a "leaning attack".

Top
 Profile  
 
Offline
 Post subject: Re: Video series: Learning from AlphaGo
Post #22 Posted: Tue Jan 02, 2018 4:38 pm 
Lives in gote
User avatar

Posts: 392
Liked others: 23
Was liked: 43
Rank: NR
gennan wrote:
I can only guess what AlphaGo and pros are thinking
right - and as i said before, your guess is better than mine - and it's normal primate behaviour to "project", ie assume that when someone else does something, they do it for the reasons we imagine we would have if we had done it - this is the basis of empathy. It's also what makes magic work, when the audience simply cannot imagine how the magician pulled off his delightful trick - and it's what enables confidence tricksters like the Jekyll Island mob and their present-day incarnations to pull off dirty tricks when their victims have no idea that a trick is being pulled, as they did in 1929, 2008, and coming soon to a wallet near you.

Ever since 1971, when i started thinking about what language means, it's been a mystery to me how people ever manage to understand even half of what somebody else says, as natural language is so elliptical and the imaginary world inside anyone's head so intricately interwoven that the probability of misunderstanding is miles bigger than the probability of understanding. Actions speak louder than words, but Alfie's actions are based on icebergs so deep that not even Ke Jie can fathom her intentions.

And although Alfie - who is no more a thinking machine than ENIAC was, but like ENIAC, better at doing sums than human computers - has thrown a spanner in the works of AI by demonstrating that OR is as handy at Go as it is at calculating ballistic missile trajectories (the reason ENIAC was funded), it's still the case that an Intelligent Plastic Machine can project his mindset onto Alfie's behaviour to make some sense of her machinations.

It's a mathematical certainty that there is a truth about Go, even if it's equally mathematically certain that we will never know what that truth is, as there aren't enough atoms in the known Universe to build a machine to calculate it - but as the same is true of chess, who cares?

Be that as it may, evolution has programmed into us a relentless curiosity to seek the truth even when we know we can never find it; it amuses us to climb mountains for no other reason than that they are there, and even when we know that like the 5 pendulums, it's an utterly pointless exercise that does more to damage the ego than to stroke it (puns intended).


gennan wrote:
I see only one response by [AlphaGo and pros] after white 4 (=move 5 at E14). When only one move is possible for move 5, I see white 4 as a forcing move
Right enough, AlfieTeach only shows E14, although AT is heavily redacted (presumably by Fan Hui). It is far from obvious to me that white 4 is forcing - the Go mountain in front of me is like a sand dune, one step up and two steps down each time Christmas comes around and a few more brain cells go down the toilet.
Attachment:
white4.png
white4.png [ 274.63 KiB | Viewed 3552 times ]

gennan wrote:
I kind of understood that you wanted to considering other options for black's move 5 to take the outside instead of the inside
black's options for 5 did pique my interest, but it wasn't my original question, which was about black 3. To restate, whereas i guessed that black 3 indicated interest in the top, that doesn't mean i am a black and white person who would insist on the top with black 5, no matter what white does.

White 4 looks to me like white also "thinks" black 3 partly intends the top, and is a move to counter that intention. But that doesn't mean to me that black 3 was purely a decoy.

gennan wrote:
perhaps you weren't even asking me a question (or perhaps you were, but I got lost somewhere).
it's ancient history now, but my question was about whether black 3 could have 2 meanings.

gennan wrote:
move 4 .. is .. a "leaning attack".
now you mention it, i realise that Swim too sees it that way, although that perspective didn't occur to me (i didn't even think about what white 4 might mean until yesterday, and it hadn't occurred to me to consider the possibility of white 4 when thinking about what black 3 could mean. My foresight of lookahead depth = 0 probably explains why Go mountain is for me a step too far).

Swim sees white 4 as an attachment to a single stone in a group (in this case, the group of just one stone: black 1) with the intention that while black defends, white can build outside strength preparatory to reducing a black moyo (in this case, the one formed by G16 and Q16).

Code:
to leaningattack (x from y)
     do for  block in x
          where (adjacent (block, border(y)) and size (block = 1)
          attach (y, block) or shoulderhit (y, block)



It's a side-issue, but AT says Alfie prefers F17 to white 4, despite her win% of it being much less. Maybe Alfie is an Alf Garnet Trump voter, choosing a second-rate but popular candidate, going with the flow, following the flock; like events in Europe, it's looking more and more like another of the JI mob's prestidigitations.

Q: What's the difference between Go and real life?
A: In real life, everything is connected.

"Who said that? The party chairman?" asked Bernard.
"Nearly right", replied Sir Humphrey, "Actually, it was Lenin".

Follow-up question: what, if anything, can AT teach me about white's invasion? at the time, i thought it was on the standard invasion point, but black went after it like a ferret; in retrospect i would say it was my losing move, and i should have prepared an escape route before jumping in so deep. i had played it as a sacrifice probe, but fighting spirit got the better of me and when a long thin straggly black group running into the centre finally made eyes, black's influence was overwhelming.
Attachment:
invade.png
invade.png [ 477.68 KiB | Viewed 3539 times ]

_________________
i shrink, therefore i swarm


This post by djhbrown was liked by: daal
Top
 Profile  
 
Offline
 Post subject: Re: Video series: Learning from AlphaGo
Post #23 Posted: Wed Jan 03, 2018 12:19 am 
Lives with ko

Posts: 143
Liked others: 45
Was liked: 33
Rank: EGF 3d
KGS: gennan
Tygem: gennan
OGS: gennan
Kaya handle: gennan
djhbrown wrote:
.. AT says Alfie prefers F17 to white 4, despite her win% of it being much less.


You are not the first to miss this, but all winrates in AT are black's winrates (as stated in the legend below the board and as I explained at the start of my video #1). So white's precentages are 46.1 for E15 and 51.1% for F17. AG prefers F17 because of that rather big difference.

My conclusion in my video #1 is that E15 a rather big mistake by white. This is surprising, because E15 has been considered a joseki move for many decades. These kind of surprises are the reason I'm making these videos.

Top
 Profile  
 
Offline
 Post subject: Re: Video series: Learning from AlphaGo
Post #24 Posted: Wed Jan 03, 2018 12:31 am 
Lives with ko

Posts: 143
Liked others: 45
Was liked: 33
Rank: EGF 3d
KGS: gennan
Tygem: gennan
OGS: gennan
Kaya handle: gennan
djhbrown wrote:
Follow-up question: what, if anything, can AT teach me about white's invasion? at the time, i thought it was on the standard invasion point, but black went after it like a ferret; in retrospect i would say it was my losing move, and i should have prepared an escape route before jumping in so deep. i had played it as a sacrifice probe, but fighting spirit got the better of me and when a long thin straggly black group running into the centre finally made eyes, black's influence was overwhelming.
Attachment:
invade.png


If the game didn't go as expected after this, it does not mean that this move was the cause. If you'd like a review, I suggest you upload the game to https://gokibitz.com

Top
 Profile  
 
Offline
 Post subject: Re: Video series: Learning from AlphaGo
Post #25 Posted: Wed Jan 03, 2018 1:36 am 
Lives in gote
User avatar

Posts: 392
Liked others: 23
Was liked: 43
Rank: NR
gennan wrote:
all winrates in AT are black's winrates
Doh! I should have guessed that a mindless moron like Alfie wouldn't know that sensible readers reasonably assume a label on a white stone says something about white :)

gennan wrote:
E15 a rather big mistake by white. This is surprising, because E15 has been considered a joseki move for many decades. These kind of surprises are the reason I'm making these videos
Hindsight is a wonderful thing, but now that you have shown us that Her Royal Alfness has pointed it out, E15 does look kind of clumsy, perhaps an overreaction to G16 and unnecessarily confrontational, in the psycho "cut before you think" mould, which seems to be how almost all my opps play. That i invariably succumb to such brutish tactics only encourages them, but the queen shows us that deftness beats brute force, which is kind of ironic, seeing as she uses brute force inside her head to come up with a delicate touch on the board. Maybe it's just that far-sighted brute force beats short-sighted belligerence.

gennan wrote:
If you'd like a review, I suggest you upload the game to https://gokibitz.com
thanks for the offer; i had another look at it and saw that yes, it wasn't that move which was the beginning of the end, but an elementary tactical blunder a few moves later... and then to make matters worse, i completely forgot that my blunder meant that the thin straggly black group wasn't eyeless at all, because it had connected by cutting me where i forgot to protect. so all in all, too dumb a game to test anyone's patience to review. and it's not as if the talk by Rudd i was watching at the same time was anything worth listening to.

What's the name of that disease where you forget things?

_________________
i shrink, therefore i swarm

Top
 Profile  
 
Offline
 Post subject: Re: Video series: Learning from AlphaGo
Post #26 Posted: Wed Jan 03, 2018 6:36 am 
Gosei

Posts: 1426
Liked others: 709
Was liked: 471
Rank: AGA 3k KGS 1k Fox 1d
GD Posts: 61
KGS: dfan
I prefer the win rate to always be from the same point of view (e.g., Black), regardless of whose turn to move it is. I hate analyzing with Leela and having to spend half a second every move figuring out how much the evaluation really changed when the win rate went from 57% to 42%. Crazy Stone does this right (from my perspective).


This post by dfan was liked by: gennan
Top
 Profile  
 
Offline
 Post subject: Re: Video series: Learning from AlphaGo
Post #27 Posted: Thu Jan 04, 2018 12:50 am 
Lives with ko

Posts: 143
Liked others: 45
Was liked: 33
Rank: EGF 3d
KGS: gennan
Tygem: gennan
OGS: gennan
Kaya handle: gennan
djhbrown wrote:
gennan wrote:
all winrates in AT are black's winrates
Doh! I should have guessed that a mindless moron like Alfie wouldn't know that sensible readers reasonably assume a label on a white stone says something about white :)


Yes, I also found this presentation choice by DeepMind a bit surprising and from comments I understand that it was counterinituitive for others as well. I was used to the Waltheri database presentation which shows the winrates for the player to move. So I had to do conversions when comparing Waltheri winrates with AT winrates for my video #2.

First I thought that DeepMind made an error against the principle of least surprise, but dfan actually prefers the AT presentation. Perhaps he has a point. Initially it is a bit surprising to many people, but when one gets used to it, it may be more convenient like this.

Top
 Profile  
 
Offline
 Post subject: Re: Video series: Learning from AlphaGo
Post #28 Posted: Thu Jan 04, 2018 2:59 am 
Dies in gote

Posts: 39
Liked others: 40
Was liked: 10
gennan wrote:
djhbrown wrote:
gennan wrote:
all winrates in AT are black's winrates
Doh! I should have guessed that a mindless moron like Alfie wouldn't know that sensible readers reasonably assume a label on a white stone says something about white :)
Yes, I also found this presentation choice by DeepMind a bit surprising
From a developer's point of view this makes a lot of sense. They chose blacks winning percentage as target variable p. Black tries to maximize p ( p->100%) and White tries to minimize it (p -> 0%). That's a standard procedure, the minimax principle.

Being kind of a developer myself (actually I'm mathematician) I didn't have a problem with that from the start, but I imagined upcoming confusion. There seems to be a principle: Everything that can be misunderstood will be misunderstood ;-)

_________________
Couch Potato - I'm just watchin'!

Top
 Profile  
 
Offline
 Post subject: Re: Video series: Learning from AlphaGo
Post #29 Posted: Thu Jan 04, 2018 7:25 am 
Lives in gote
User avatar

Posts: 392
Liked others: 23
Was liked: 43
Rank: NR
Baywa wrote:
From a developer's point of view this makes a lot of sense.
Yes, indeed, convenience for the programmer and les users peuvent manger de la brioche. Typical, bloody typical.

One more for the Hall of Shame
Attachment:
hci.png
hci.png [ 185.72 KiB | Viewed 3393 times ]


Top
 Profile  
 
Offline
 Post subject: Re: Video series: Learning from AlphaGo
Post #30 Posted: Sat Jan 06, 2018 12:50 pm 
Lives with ko

Posts: 143
Liked others: 45
Was liked: 33
Rank: EGF 3d
KGS: gennan
Tygem: gennan
OGS: gennan
Kaya handle: gennan
I just finished episode #3 about the Chinese Opening: https://www.youtube.com/watch?v=qv0gg0RJpZQ


This post by gennan was liked by 3 people: Baywa, Bill Spight, sybob
Top
 Profile  
 
Offline
 Post subject: Re: Video series: Learning from AlphaGo
Post #31 Posted: Sun Jan 21, 2018 12:56 pm 
Lives with ko

Posts: 143
Liked others: 45
Was liked: 33
Rank: EGF 3d
KGS: gennan
Tygem: gennan
OGS: gennan
Kaya handle: gennan
I just finished episode #4: 5-space extensions are wrong

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

Top
 Profile  
 
Offline
 Post subject: Re: Video series: Learning from AlphaGo
Post #32 Posted: Tue Feb 06, 2018 1:43 pm 
Lives with ko

Posts: 143
Liked others: 45
Was liked: 33
Rank: EGF 3d
KGS: gennan
Tygem: gennan
OGS: gennan
Kaya handle: gennan
Learning from AlphaGo #5: refuting a greedy joseki

https://www.youtube.com/watch?v=5ouxL3G_tIo

Top
 Profile  
 
Offline
 Post subject: Re: Video series: Learning from AlphaGo
Post #33 Posted: Wed Feb 07, 2018 3:11 am 
Oza
User avatar

Posts: 2494
Liked others: 1280
Was liked: 1110
I learned something from your last video that I hadn't previously understood. It is your explanation starting around 3:35 pointing out that the kikashi doesn't lose anything and can be treated lightly, because even if black swallows up the white stone, the loss is compensated by the fact that b had been forced to play inside his own territory. Thanks! Keep up the good work, I like your videos a lot.

BTW, why does b always have such a miserable winrate, and is this opening winrate of 47% reflected in the overall results of alphago vs. alphago? If so, it seems a clear indication that komi is wrong, no?

_________________
Go? It's like sharing two bowls of cookies, but one person gets all of them.


This post by daal was liked by: gennan
Top
 Profile  
 
Offline
 Post subject: Re: Video series: Learning from AlphaGo
Post #34 Posted: Wed Feb 07, 2018 5:38 am 
Lives with ko

Posts: 243
Location: Spain
Liked others: 151
Was liked: 24
Rank: Low
GD Posts: 10
daal wrote:
BTW, why does b always have such a miserable winrate, and is this opening winrate of 47% reflected in the overall results of alphago vs. alphago? If so, it seems a clear indication that komi is wrong, no?

AlphaGo uses 7.5 komi and Chinese rules. 7 komi (breaking ties with a button) would probably be ideal.

_________________
Sum ergo non ero.

Top
 Profile  
 
Offline
 Post subject: Re: Video series: Learning from AlphaGo
Post #35 Posted: Wed Feb 07, 2018 5:53 am 
Judan

Posts: 6189
Location: Cambridge, UK
Liked others: 354
Was liked: 3340
Rank: UK 4 dan
KGS: Uberdude 4d
OGS: Uberdude 7d
Yes, AG thinks white is better on the empty board with 7.5 komi. But I don't think that 47% number actually means black wins 47% of the million self-play games they've done, I understand it as an ill-defined goodness metric rather than a real probabilty, see discussion.

If komi was 6.5 (does that even make sense with Chinese, does it need to change in 2s?) it could be black's "win rate" is 55%, so more lopsided than with 7.5; thus 7.5 might be the non-drawing komi which is closest to even which is probably what you want*.

* Unless you go crazy and make a virtual fractional komi: say if 6.5 komi black wins 55% (+5%) and 7.5 komi 47% (-3%), therefore let's have 7 komi and if it's a draw on the board we assign win to black in 3/8 of cases and white in 5/8 of cases so komi is kinda like 7 and one eighth, where the 50% would be on a linear interpolation between our 2 data points.

Top
 Profile  
 
Offline
 Post subject: Re: Video series: Learning from AlphaGo
Post #36 Posted: Wed Feb 07, 2018 5:59 am 
Lives in gote
User avatar

Posts: 392
Liked others: 23
Was liked: 43
Rank: NR
luigi wrote:
[...komi
heading off-topic here, so let's take it there:
viewtopic.php?p=227322#p227322

_________________
i shrink, therefore i swarm

Top
 Profile  
 
Offline
 Post subject: Re: Video series: Learning from AlphaGo
Post #37 Posted: Wed Feb 07, 2018 6:33 am 
Lives with ko

Posts: 243
Location: Spain
Liked others: 151
Was liked: 24
Rank: Low
GD Posts: 10
Uberdude wrote:
7 komi and if it's a draw on the board we assign win to black in 3/8 of cases and white in 5/8 of cases so komi is kinda like 7 and one eighth

Well, with the button, if it's a draw on the board, each color is assigned the win in 1/2 of cases. :)

_________________
Sum ergo non ero.

Top
 Profile  
 
Offline
 Post subject: Re: Video series: Learning from AlphaGo
Post #38 Posted: Wed Feb 07, 2018 12:27 pm 
Lives with ko

Posts: 143
Liked others: 45
Was liked: 33
Rank: EGF 3d
KGS: gennan
Tygem: gennan
OGS: gennan
Kaya handle: gennan
It seems that 47% winrate for black with an empty board at 7.5 komi is even optimistic. From the 50 published self-play games Master vs Master, black only won 12 (24%). Also, in AlphaGo's opening database, you gradually see black's winrate drop as the game progresses, even when black plays the best moves according to AlphaGo. Usually black's winrate drops to about 43% by move 30. So as the game progresses, AlphaGo seems to become more certain that 7.5 komi is too much.

I think that mathematically perfect komi should be an integer. And because of the increasingly bad odds that black seems to have in AlghaGo's evaluation and self-play game results (using 7.5 komi), my guess would be that perfect komi is 6 rather than 7.

But for human (imperfect) games, I prefer a komi that prevents jigo. So perhaps 6.5 komi would be a fair approximation with Japanese rules. With Chinese rules, almost all games end with an odd score difference on the board. So in practice, black needs 7 points more on the board to win with 5.5, 6 or 6.5 komi. But black needs 9 points more on the board to win with 7.5 komi. (note that even though even score differences on the board are rare with Chinese rules, it is possible that the perfect game has it).

Top
 Profile  
 
Offline
 Post subject: Re: Video series: Learning from AlphaGo
Post #39 Posted: Wed Feb 07, 2018 1:04 pm 
Lives with ko

Posts: 243
Location: Spain
Liked others: 151
Was liked: 24
Rank: Low
GD Posts: 10
gennan wrote:
It seems that 47% winrate for black with an empty board at 7.5 komi is even optimistic. From the 50 published self-play games Master vs Master, black only won 12 (24%). Also, in AlphaGo's opening database, you gradually see black's winrate drop as the game progresses, even when black plays the best moves according to AlphaGo. Usually black's winrate drops to about 43% by move 30. So as the game progresses, AlphaGo seems to become more certain that 7.5 komi is too much.

That's only natural. If AlphaGo played perfectly, it would only display 0% and 100% win rates. It gets closer to perfection as the game progresses (of course, it's never going to display 50% right before passing).

Also, the 50 published games are cherry-picked.

_________________
Sum ergo non ero.

Top
 Profile  
 
Offline
 Post subject: Re: Video series: Learning from AlphaGo
Post #40 Posted: Wed Feb 07, 2018 1:49 pm 
Lives with ko

Posts: 143
Liked others: 45
Was liked: 33
Rank: EGF 3d
KGS: gennan
Tygem: gennan
OGS: gennan
Kaya handle: gennan
luigi wrote:
gennan wrote:
It seems that 47% winrate for black with an empty board at 7.5 komi is even optimistic. From the 50 published self-play games Master vs Master, black only won 12 (24%). Also, in AlphaGo's opening database, you gradually see black's winrate drop as the game progresses, even when black plays the best moves according to AlphaGo. Usually black's winrate drops to about 43% by move 30. So as the game progresses, AlphaGo seems to become more certain that 7.5 komi is too much.

That's only natural. If AlphaGo played perfectly, it would only display 0% and 100% win rates. It gets closer to perfection as the game progresses (of course, it's never going to display 50% right before passing).

Also, the 50 published games are cherry-picked.


Do you think that DeepMind cherry-picked mostly white wins, but in fact unpublished wins had an even color distibution?

Suppose that AlphaGo became so strong that it would evaluate 0% or 100% black winrate for an empty board with 7.5 komi. Will it be 0% or 100%?

My guess is it would be 0%.

But with an integer komi, there can be a komi value where perfect players would always get a jigo. The correct evaluation of an empty board with that komi value would be 50%.

Top
 Profile  
 
Display posts from previous:  Sort by  
Post new topic Reply to topic  [ 66 posts ]  Go to page Previous  1, 2, 3, 4  Next

All times are UTC - 8 hours [ DST ]


Who is online

Users browsing this forum: No registered users and 1 guest


You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot post attachments in this forum

Search for:
Jump to:  
Powered by phpBB © 2000, 2002, 2005, 2007 phpBB Group