influence concensus map

[dfan]

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.

[jaca]

I am not sure what you mean by "end of each line".

A0treesearch.png (40.11 KiB) Viewed 12038 times

https://www.nature.com/articles/nature24270

[sorin]

I am not sure what you mean by "end of each line".

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".

[jaca]

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.

[jaca]

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:

huh.png (398.72 KiB) Viewed 11942 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.

[Vio]

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

[Knotwilg]

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".

[jaca]

..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
https://www.lifein19x19.com/viewtopic.p ... 67#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!

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,

[Vio]

Insightful answer, thx I'll check sometimes the arguments in the links provided

[Aidoneus]

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!

[jaca]

All About Thickness by Ishida Yoshio

Kiseido sell an e-book download of All About Thickness
http://www.kiseidodigital.com/kdb1-toc.html

Ishida was called "the computer" by his peers, so who better to guide us to a better understanding of influence!

So i've made a start at learning what Ishida has to say and sharing my understanding of his wisdom in the first of what i hope will be a series of videos that look at what influence is and how it might be computed. https://youtu.be/OfL5XSrYuLo Update: Yishn has told me that he finds the video narrator's accent hard to understand, so i will make the next one with a different voice.

here's what i believe could be a more intuitively meaningful way of displaying Lz's win%. Each evaluated move is coloured by 2 rings, the thick inner ring is win% and the thin outer (harmonic) ring is delta(win%) which is the difference between black's win% for that move compared to black's win% just before white played her last move. The bulls-eye of the move is the colour of the player on move, in this case, black. i have also elided Sabaki's 'last move' mark, as (a) in Go, which move was last is only of historical interest (historians can easily find out by strolling back and forth, or by looking in the the game tree subwindow), and (b) its strong colour conflicts with bullseye colour.

ex1recolour.png (364.95 KiB) Viewed 6957 times

Explanation of the colour scheme rationale is provided in a GitHub message
https://github.com/SabakiHQ/Sabaki/issu ... -455859627
win% is adjusted to take into account handicap before colouring, so what Sabaki says is 92%, the handicap-adjusted scheme says is about 52%. Here, all Laizy's top moves increase black's win%, which remains within the olive zone.
P7's delta ring is a tiny bit lighter than all the others, which are all less than 1% higher than P7's (i forget the exact numbers).

------------------------
Opinions expressed in this article are not necessarily endorsed by its publisher.

[Vio]

Tomorrow we will advise to not read that kind of book unless you have a good understanding on how AI play go. The reverse as what we advised yesterday.

Some stronger players did already advise me on this in my teachings...

[jaca]

Tomorrow we will advise The reverse as what we advised yesterday.

Every woman has a right to change her mind, and every man has a right to try his level best to change it for her.

Each year, there is an annual mind-changing festival in Paris (= per Isis) at the Spring Equinox of Oestre (Isis), when a young man's fancy turns to thoughts of love and a young woman's head is turned toward his pocketbook by seamstress queans eager to urge her to throw away that old dress and get him to buy her a new one.

Laizy's great-grandmother was born in Paris https://xkcd.com/1263/, and the influence her daughter exerted over the minds of Dedicated Followers of Fashion https://www.youtube.com/watch?v=QA5gJ0hZpCc cannot be overemphasised, for She Who Must Be Obeyed - the Alpha and Omega Queen of Queans of the Go catwalk - is no False Prophet but a genuine Oracle with 20-20 vision extending far over the rainbow and into the wide blue yonder, so any swine who does not heed Her pearls of wisdom is a ridiculous Bottom.

The only problem is: How the hell can we make any sense of what She says? Aye, there's the rub, for She doesn't say a damn thing, she only does (mares eat does and does eat does, but little lambs eat Ivy - poor old Ivy, doomed for all eternity to perpetually climb like Sisyphus up the wall of thickness with nowhere to Go!)

Sure, anyone can drive themself into a 3-3 corner without thinking about it, and anyone can do a double jump along the 4th line because that's what Laizy's grandmother liked to do; but surely we can learn more from Her than that??!

[jaca]

...surely we can learn more from Laizy than just where to play and principal variations?

i have a dream...

there is no finer commentator on Laizy than Michael Redmond, but there's only one of him and he has better things to do than comment on your and my games.

so my dream is Rbot, a Redmond-inspired AI that can tell me what's going on in a position and why such-and-such is a good move, or not, as the case may be.

to that end, i took another look at one of his commentaries to see if there were any themes that could be captured and expressed in a kind of "Go-logic".

Screenshot from 2019-01-25 07-04-23.png (141.86 KiB) Viewed 7037 times

i haven't got very far, but have seen a couple of things that encourage me to think that Rbot might not be a pipe-dream, even if it's gonna take more than a weekend to write the code for.

With the help of auto-generated captions, i made the following notes on what Mr Mike said - i'm not looking so much at the variations he talks about but the language in which he talks about them. i've highlighted the key descriptive qualities that could be computable, and will explain how one of them maybe could be computed:

black: zero; white:master

1.12 [we will later see] black: groupN (big, about to die, floating around in the middle of white's moyo)
1.17 black groupN (not urgent)

game start

1.31 black (jump into 3-3)
1.59 if (opp starpoint) then invade(3-3)
2.09 upperleft_corner(even result)

2.31 [upper left joseki]: crawl (timing can change: if outer player has support in centre of side then inner player can play hane-tsugi)
2.54 eg. if white crawls now [G17] then black hane [B16] and connect [B17] leaves white slightly overconcentrated [on left side]

"slightly overconcentrated" - what does that mean? Well, it could mean that the influence emanating from white's upper wall flows over the marked point,

so that if white already had a stone there, white would have wasted that stone - i.e either she shouldnt have played it there in the first place, or she shouldnt crawl once more in the upperleft joseki because her influence radiations would then overlap too much.

overconcentration is inefficient. an influence map algorithm that actually worked properly would enable Rbot to evaluate the influence-efficiency of a move.

all Rbot would need to do is compute the influence map and measure how much the influences from the wall and a white stone on the marked point would overlap.

PS yes, yes, i know there are plenty of Goers who will think this sounds like old-fashioned "Expert Systems" which never worked and never could work, but i want to claim that those that were produced way back then were only scratching the surface and the folks back home initially expected too much of them too soon, and like bipolar people, swung too far the other way from love to hate.

[Bill Spight]

PS yes, yes, i know there are plenty of Goers who will think this sounds like old-fashioned "Expert Systems" which never worked and never could work, but i want to claim that those that were produced way back then were only scratching the surface and the folks back home initially expected too much of them too soon, and like bipolar people, swung too far the other way from love to hate.

It is not true that expert systems never worked. For instance, one medical expert system gave better initial diagnoses (that is, before tests were performed) than general practitioners. Also in the medical field, MDs were able to learn from an expert system and do better by themselves than either they or the system alone had done before.

It is going to take some time, but I think that the AI era will fairly quickly see humans playing much better. Even if they can't say why.