It is currently Sun Jul 21, 2019 12:50 pm

All times are UTC - 8 hours [ DST ]




Post new topic Reply to topic  [ 50 posts ]  Go to page Previous  1, 2, 3  Next
Author Message
Offline
 Post subject: Re: influence concensus map
Post #21 Posted: Tue Jan 08, 2019 11:26 am 
Dies with sente

Posts: 119
Liked others: 30
Was liked: 27
Rank: EGF 3d
KGS: gennan
Tygem: gennan
OGS: gennan
Kaya handle: gennan
Those influence maps look pretty, but I don't really understand the purpose. To me, they look too crude to be useful for anything.
How does it help to detect who is leading or which areas are big and / or urgent?

Top
 Profile  
 
 Post subject: Re: influence concensus map
Post #22 Posted: Tue Jan 08, 2019 2:42 pm 
Is there even such a thing as a sphere of Influence? If so, what is it? What does it mean?

Are all pretty things air-headed? Many do seem to be, but maybe looks are deceiving, for if you've got good looks, you can push a lot of buttons and pull a lot of strings and have an easy life without having to do much work - so maybe they're smarter than you think! I daresay Charles Darwin would have an opinion on this subject.

Physicists are in love with the beautiful equation - the simpler it is, the more beautiful. One of the most famous equations of all is exp(ixpi) = -1 although what that really means i'm not quite sure; does it have something to do with spheres?

John Keats reckoned that Beauty was Truth, and Truth, Beauty. https://www.poetryfoundation.org/poems/ ... recian-urn

Was he wrong?


---------------
Know that this worldly life is no more than playing games, and boasting among you, .. It is like abundant rain that produces plants and pleases the disbelievers. But then the plants turn into useless hay, and are blown away by the wind. (Al-Hadeed 57:20)

Top
  
 
Offline
 Post subject: Re: influence concensus map
Post #23 Posted: Tue Jan 08, 2019 11:42 pm 
Lives in sente

Posts: 1121
Location: Earth
Liked others: 361
Was liked: 183
When I play I dont care about influence. My friends care about influence. I care about good moves. I win.

;-)


This post by Gomoto was liked by: sorin
Top
 Profile  
 
Offline
 Post subject: Re: influence concensus map
Post #24 Posted: Wed Jan 09, 2019 1:35 am 
Judan

Posts: 5942
Location: Cambridge, UK
Liked others: 338
Was liked: 3187
Rank: UK 4 dan
KGS: Uberdude 4d
OGS: Uberdude 7d
Speaking of influence, here's a nice example of LeelaZero invading deeper than a superficial judgement of what you might think is reasonable (if I played it I could well imagine a teacher telling me off for going too deep) based on the surrounding influence, but due to weakness in the influence it's actually surprisingly hard to punish.

viewtopic.php?f=18&t=16198&p=240382#p240382


This post by Uberdude was liked by: Vio
Top
 Profile  
 
Offline
 Post subject: Re: influence concensus map
Post #25 Posted: Wed Jan 09, 2019 1:15 pm 
Lives in sente
User avatar

Posts: 1306
Location: Ghent, Belgium
Liked others: 184
Was liked: 582
Rank: Bel 2d KGS 3d TG 4d
KGS: Artevelde
Tygem: Knotwilg
Here I discuss the part of the OP not related to influence: moves 19, 20, 21 and variations.


Top
 Profile  
 
 Post subject: Re: influence concensus map
Post #26 Posted: Sat Jan 12, 2019 2:39 pm 
Gomoto wrote:
When I play I dont care about influence. My friends care about influence. I care about good moves. I win.;-)
LZ doesn't care about influence either... So she isn't one of Gomoto's friends?? i thought everybody loved LZ!

The human brain learns by forming new synapses, and modifying existing ones, a process not dissimilar to the way a DCNN modifies its weights.

LZ figures out how to modify its DCNN weights by reading all the way to the end of the game, seeing who won, and backing up that information in the context of all the other variations in the forest it has looked at.

Even if Gomoto doesn't read all the way to the end of the game at every move, he at least later on finds out whether he won or lost in actual play (a forest of width = 1).

So if he learns anything from his wins and losses, Gomoto could be said to learn Go in the same kind of way that LZ does, because (a) they both learn by modifying the strengths of synapses, and (b) they both don't care about influence.

https://github.com/gcp/leela-zero/blob/ ... #compiling says:
"Recomputing the AlphaGo Zero weights will take about 1700 years on commodity hardware."

Let F be the width of the forest of variations LZ explores.
Batman always wins - So, because he always wins against his friends, Gomoto :batman: must be (1700 x F) years old, or he has a brain (1700 x F) times bigger than mine.

Wait a minute! :batman: only cares about good moves, and we all know that the only good moves are the ones LZ makes, so :batman: is LZ!!!!!!


gennan wrote:
Those influence maps.. they look too crude to be useful for anything.
How does it help to detect who is leading or which areas are big and / or urgent?

An Example of Influence Map Utility: Visualising Tradeoffs between Obedient Gote and Wishful Tenuki


here are two rows of pictures of different white responses to black L4, to show the tradeoff between:
top row = blocking in gote at M4 to save a couple of stones and prepare the way for a later invasion/reduction of black's bottom right corner
bottom row = taking sente to gain influence elsewhere, eg D10
Attachment:
leela226.png
leela226.png [ 457.37 KiB | Viewed 1132 times ]

Bill Spight wrote:
I do think that an influence map produced by having a top bot play itself from the current position 1,000,000 times could be quite educational, however
The Leela maps in the above picture are her Moyo markup. Her maps after thinking are a bit different from those before thinking. For M4, Leela changes her mind a little bit after looking at 2.5 million nodes... she made my cpu get hot, so for the second run (D10) i took a snapshot after only 700,000 nodes - short of a million, but the map was pretty stable by then; you can see it is more than a bit different from her first guess (which has black's lower group smothered by white's massive moyo!).

i like the look of what Influencie produces, as it corresponds with my own intuition - although this example might be a touch optimistic about connection strength of long jumps? Influencie lets you play around with Stone influence power; i may have had it turned up too high.

the 3 different mappers agree on a lot of things, but there are a few differences.

Influencie wins the beauty competition by a mile, and the size of her left black shadow agrees with the guess i made during my game review with W, so i am biased like hell in her favour (Oh, i remember now, i twiddled with her Stone Influence control until she did! :oops:) - but the other two may have sounder judgement.

Top
  
 
Offline
 Post subject: Re: influence concensus map
Post #27 Posted: Sat Jan 12, 2019 6:15 pm 
Honinbo

Posts: 8593
Liked others: 2514
Was liked: 2962
Who says LZ doesn't care about influence? If LS cares about anything it cares about influence, that is, about the effect of stones on other stones and empty points.

_________________
There is one human race.
----------------------------------------------------

The Adkins Principle:

At some point, doesn't thinking have to go on?

— Winona Adkins

Top
 Profile  
 
 Post subject: Re: influence concensus map
Post #28 Posted: Sat Jan 12, 2019 8:01 pm 
Bill Spight wrote:
If LS cares about anything it cares about influence, that is, about the effect of stones on other stones and empty points.
To us humans, it sure looks like she does, especially when she makes a fool of Monte-Carlo by rushing in like an angel where devils fear to tread.

But in fact - and you can check this with her programmers - when playing, LZ only thinks about who wins at the ends of each of her 1,000,000 imagined variations per move.

Just like :batman:, her eyes are constantly fixed solely upon the end of time, whilst her mind's eye interprets where she is now in terms of what she sees at the end of time.

But she never never never ever thinks about influence - or any other high-level cognitive concept, for that matter

... although i guess you could say that convolutions compute abstractions, so i can only say with confidence that she doesn't think about any other high-level human cognitive concept, which is going to make talking to her a bit tricky -

unless it turns out that some convolutions do have a correlation with human perception interpretation - things like being able to draw a line around an object in a scene, which face recognition and melanoma classification software is now able to do

- but at the moment, i don't see how LZ could employ the pixel abstraction technique to draw a line around a Go group, because what makes a Go group a group is not just its surface geometry.

Top
  
 
Offline
 Post subject: Re: influence concensus map
Post #29 Posted: Sat Jan 12, 2019 8:21 pm 
Honinbo

Posts: 8593
Liked others: 2514
Was liked: 2962
jaca wrote:
Bill Spight wrote:
If LS cares about anything it cares about influence, that is, about the effect of stones on other stones and empty points.
To us humans, it sure looks like she does, especially when she makes a fool of Monte-Carlo by rushing in like an angel where devils fear to tread.

But in fact - and you can check this with her programmers - when playing, LZ only thinks about who wins at the ends of each of her 1,000,000 imagined variations per move.


What you have described is a MCTS player. LeelaZero has taken the next step.

_________________
There is one human race.
----------------------------------------------------

The Adkins Principle:

At some point, doesn't thinking have to go on?

— Winona Adkins

Top
 Profile  
 
 Post subject: Re: influence concensus map
Post #30 Posted: Sat Jan 12, 2019 10:24 pm 
Bill Spight wrote:
What you have described is a MCTS player. LeelaZero has taken the next step.
Both A(Leela) and A0(LZ) read out to the end of each line, and then back up whether they saw a win or a loss; from that point of view, they are the same.

The key difference between them is that in both her training and playing, A0 does not consider randomly-generated moves - that's what makes her 7-league boots better than A, and what gives her a different playing-style, one that corresponds more to human received wisdom, except for her remarkable ability to invade, not to mention being able to read wider and deeper than all of China could working together if every single one of them were as strong as Ke Jie! Yes, there are 2 billion people in China, but not even he can read 300 moves deep.

i don't know abut you, but i'm not like Gomoto :batman: - i do try to think about influence, even if i don't get it quite right and/or don't use it in the best way possible. Like Manuel, "I try, Mister Fawlty - I try!!".

Top
  
 
Offline
 Post subject: Re: influence concensus map
Post #31 Posted: Sun Jan 13, 2019 6:38 am 
Gosei

Posts: 1373
Liked others: 672
Was liked: 444
Rank: AGA 3k KGS 1k
GD Posts: 61
KGS: dfan
jaca wrote:
Both A(Leela) and A0(LZ) read out to the end of each line, and then back up whether they saw a win or a loss; from that point of view, they are the same.

I am not sure what you mean by "end of each line". AlphaGo Zero etc. do not read out to the end of the game (rollouts, in MCTS parlance). In fact, they can play at a high dan level without doing any reading at all.

Top
 Profile  
 
 Post subject: Re: influence concensus map
Post #32 Posted: Sun Jan 13, 2019 2:27 pm 
dfan wrote:
I am not sure what you mean by "end of each line".
Attachment:
A0treesearch.png
A0treesearch.png [ 40.11 KiB | Viewed 1074 times ]
https://www.nature.com/articles/nature24270

Top
  
 
Offline
 Post subject: Re: influence concensus map
Post #33 Posted: Sun Jan 13, 2019 4:15 pm 
Lives in gote

Posts: 334
Liked others: 282
Was liked: 170
jaca wrote:
dfan wrote:
I am not sure what you mean by "end of each line".
Attachment:
A0treesearch.png
https://www.nature.com/articles/nature24270


Leaf node in the diagram you showed just means "unexplored node", as in no follow-up moves were looked at yet. It does not mean "end of game position".

_________________
Sorin - 361points.com


Last edited by sorin on Sun Jan 13, 2019 6:48 pm, edited 1 time in total.

This post by sorin was liked by 2 people: Bill Spight, dfan
Top
 Profile  
 
 Post subject: Re: influence concensus map
Post #34 Posted: Sun Jan 13, 2019 5:59 pm 
sorin wrote:
Leaf node.. does not mean "end of game position".
Thank you to Bill, dfan and Sorin for your corrections.

To read, or not to read,
That is the question

Whether 'tis nobler in the mind to sacrifice,
Or to take arms against a pincer
And by opposing, mock it.

To die - to read no more;
And by a death to say we end
The eye-steal and the thousand natural cuts
That stones are heir to.

'Tis a consummation devoutly to be wish'd.

To die, to read;
To read, perchance to dream—

Ay, there's the rub.

For in that read of death,
What aji may come when we have shuffled off this mortal coil
Must give us pause—
There's the respect that makes calamity of clinging on.

For who would bear the semeais of time,
Th'oppressor's wrong, the proud Dan's contumely,
The pangs of dispriz'd love, the law's delay,
The insolence of rank, and the spurns
That patient merit of th'unworthy takes,

When he himself might his quietus make
With a bare tenuki?

Who would influence bear,
To grunt and sweat under a weary life?

But that the dread of nothing after death,
The undiscovere'd country, from whose bourn
No traveller returns,
Puzzles the will,

And makes us rather bear those ills we have
Than fly to others that we know not of

Thus, conscience,
Doth make cowards of us all

And so, the native hue of resolution
Is sickled o'er with the pale cast of thought,

And Leelas of great pitch and moment
With this regard, their currents turn awry

And lose the name of action.

Top
  
 
 Post subject: Re: influence concensus map
Post #35 Posted: Wed Jan 16, 2019 3:57 pm 
Everyone makes decisions with some combination of intuition and analysis. Some are more intuitive; some more analytical.
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4519675/

It's probably the case that the best Go players are more analytical than intuitive, as reading isn't something you can afford to not do.

But what to read??

i only ever speak for myself, and everything i say is only my opinion, and in this case my opinion is that i would like some help in figuring out what to read.

i could just ask a machine to just tell me where to play and hope i could make some sense of its choice, but having already seen many examples of the wisdom of our new overlords, with hand on heart i must confess i haven't after all learned much from them, for if i had, my playing strength would have improved from that experience, and it hasn't.

So, although the idea of influence maps has so far attracted only negative comment, i am pressing on with my lonely journey into the unknown, aware of the old homily about the journey being more pleasurable than the anticlimax of the arrival:

Since leela-zero is the best available machine since the automatic bread-slicer, any influence concensus map should take into account her opinion.

With that in mind, i bit the bullet and stumbled through a quagmire of ifs and buts to get leela-zero onto my ancient peecee, and produced this picture of Sabaki's Toggle Analysis of leela's opinion:
Attachment:
huh.png
huh.png [ 398.72 KiB | Viewed 978 times ]

but now i'm confused, because it looks to me like she's saying she would like to play on places where there are already stones! Help!! Is it the case that she's analysing an empty board instead of the one i gave her?

Can anyone tell me which button i should have pressed instead of F4? I tried pressing Toggle Guess Mode before F4, but that didn't make any difference.

Top
  
 
Offline
 Post subject: Re: influence concensus map
Post #36 Posted: Thu Jan 17, 2019 10:27 am 
Dies with sente

Posts: 70
Liked others: 17
Was liked: 21
Rank: cn 3d
OK tell me if I am wrong but
We could put an influence value for each point based on how often an opposite colour stone is not played on it in a valuable playout?
Although i guess it would be a bit heavy parallel process

Top
 Profile  
 
Offline
 Post subject: Re: influence concensus map
Post #37 Posted: Thu Jan 17, 2019 12:26 pm 
Lives in sente
User avatar

Posts: 1306
Location: Ghent, Belgium
Liked others: 184
Was liked: 582
Rank: Bel 2d KGS 3d TG 4d
KGS: Artevelde
Tygem: Knotwilg
Vio wrote:
OK tell me if I am wrong but
We could put an influence value for each point based on how often an opposite colour stone is not played on it in a valuable playout?
Although i guess it would be a bit heavy parallel process


Since "your opponent's key point is your own" this would not distinguish between either influence.
Moreover, I believe it's wrong to "play close to strength" but not to "play close to influence".

Top
 Profile  
 
 Post subject: Re: influence concensus map
Post #38 Posted: Thu Jan 17, 2019 1:14 pm 
Vio wrote:
..We could put an influence value for each point based on how often an opposite colour stone is not played on it in a valuable playout?
the logic behind your idea is harmonious with:
https://www.reddit.com/r/cbaduk/comment ... influence/
and
viewtopic.php?p=240267#p240267

It could be argued that what we see in front of our eyes right now is just a temporary illusion, since the real meaning of a present situation is how it's going to play out.

LZ is more prescient than Lady MacBeth, Nostradamus, horoscopes, the Oracle of Delphi, the writer of the Book of Revelations, crystal ball gazers, Ray Kurzweil and soothsayers of the Middle Ages, none of whom can see into the future any better than a stockmarket speculator without insider knowledge, and only tell people what they want to hear (or are afraid of!) :)

Newton did a better job of predicting where things will go - basically the same way as they are going now, until something like gravity gets in the way.

Ay, there's the rub,...

... for in general, there's a big difference between who owns what now and who will end up owning what in the end, as there are all sorts of reversals and exchanges that could take place in the meantime.

Perhaps it would be worth taking a step back and re-examining the basic concept of influence.

To me, black has influence over a point right now if it's not in white's interest to go there right now, regardless of who might end up owning it after another 100 moves. I think that's what Knotwig and the proverb "stay away from thickness" are saying.

So if that's what [current] influence means, then maybe LZ doesn't have anything meaningful to tell us after all!?

Or does she?... Influence calculations that don't take into account life and death are fundamentally so wide of the mark that they are, to use Gell-Mann's aphorism, so bad they are not even wrong! Could LZ be helpful to discover the life and death status of existing stones?
https://github.com/SabakiHQ/deadstones

Furthermore, if my definition of what influence means is correct, LZ could tell us whether it's not in white's interest to go there right now by seeing whether black's win% would increase if white did so. ... Oh, wait a sec, it's not that simple, as there are plenty of other points that it's also not in white's interest to play, such as filling in her own eyes!

Vio wrote:
Although i guess it would be a bit heavy parallel process
yes, but it wouldn't be any heavier than just asking LZ to choose a move, as the additional computation to just update numbers on a static map is negligible,


Last edited by jaca on Thu Jan 17, 2019 1:48 pm, edited 3 times in total.
Top
  
 
Offline
 Post subject: Re: influence concensus map
Post #39 Posted: Thu Jan 17, 2019 1:23 pm 
Dies with sente

Posts: 70
Liked others: 17
Was liked: 21
Rank: cn 3d
Insightful answer, thx I'll check sometimes the arguments in the links provided

Top
 Profile  
 
Offline
 Post subject: Re: influence concensus map
Post #40 Posted: Thu Jan 17, 2019 2:54 pm 
Lives in gote

Posts: 602
Location: Indiana
Liked others: 114
Was liked: 174
I don't have anything useful to contribute, but I just noticed this thread while reading All About Thickness by Ishida Yoshio, which extensively employs influence maps. Please carry on!


This post by Aidoneus was liked by: Bill Spight
Top
 Profile  
 
Display posts from previous:  Sort by  
Post new topic Reply to topic  [ 50 posts ]  Go to page Previous  1, 2, 3  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