Re: [Computer-go] AlphaGo Zero self-play temperature

[uurtamo .]

It's interesting to leave unused parameters or unnecessary
parameterizations in the paper. It telegraphs what was being tried as
opposed to simply writing something more concise and leaving the reader to
wonder why and how those decisions were made.

s.

On Nov 7, 2017 10:54 PM, "Imran Hendley"  wrote:

> Great, thanks guys!
>
> On Tue, Nov 7, 2017 at 1:51 PM, Gian-Carlo Pascutto  wrote:
>
>> On 7/11/2017 19:07, Imran Hendley wrote:
>> > Am I understanding this correctly?
>>
>> Yes.
>>
>> It's possible they had in-betweens or experimented with variations at
>> some point, then settled on the simplest case. You can vary the
>> randomness if you define it as a softmax with varying temperature,
>> that's harder if you only define the policy as select best or select
>> proportionally.
>>
>> --
>> GCP
>> ___
>> Computer-go mailing list
>> Computer-go@computer-go.org
>> http://computer-go.org/mailman/listinfo/computer-go
>>
>
>
> ___
> Computer-go mailing list
> Computer-go@computer-go.org
> http://computer-go.org/mailman/listinfo/computer-go
>
___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] AlphaGo Zero self-play temperature

[Imran Hendley]

Great, thanks guys!

On Tue, Nov 7, 2017 at 1:51 PM, Gian-Carlo Pascutto  wrote:

> On 7/11/2017 19:07, Imran Hendley wrote:
> > Am I understanding this correctly?
>
> Yes.
>
> It's possible they had in-betweens or experimented with variations at
> some point, then settled on the simplest case. You can vary the
> randomness if you define it as a softmax with varying temperature,
> that's harder if you only define the policy as select best or select
> proportionally.
>
> --
> GCP
> ___
> Computer-go mailing list
> Computer-go@computer-go.org
> http://computer-go.org/mailman/listinfo/computer-go
>
___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] AlphaGo Zero Loss

[Wesley Turner]

I can only speculate, but I see two advantages to using MSE:

* MSE accomodates games that have more than just win/loss.  One of
  AlphaGo Zero's goals (I'm extrapolating from the paper) was to develop
  a system that was easy to apply to domains other than go.

* It can be used with TD-lambda-like schemes that don't propagate
  information all the way back from a terminal position.  The AlphaGo
  team may have chosen MSE for this reason while still experimenting
  with different methods for learning and then never revisited their
  decision.

For a network that plays go and is trained as AlphaGo Zero is, I don't
see an a priori advantage for MSE over log loss, either.

On Tue, 7 Nov 2017, Petr Baudis wrote:


 Hi!

 Does anyone knows why the AlphaGo team uses MSE on [-1,1] as the value
output loss rather than binary crossentropy on [0,1]?  I'd say the
latter is way more usual when training networks as typically binary
crossentropy yields better result, so that's what I'm using in
https://github.com/pasky/michi/tree/nnet for the time being, but maybe
I'm missing some good reason to use MSE instead?

 Thanks,

--
  Petr Baudis, Rossum
  Run before you walk! Fly before you crawl! Keep moving forward!
  If we fail, I'd rather fail really hugely.  -- Moist von Lipwig
___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] AlphaGo Zero Loss

[Gian-Carlo Pascutto]

On 7/11/2017 19:08, Petr Baudis wrote:
> Hi!
> 
> Does anyone knows why the AlphaGo team uses MSE on [-1,1] as the
> value output loss rather than binary crossentropy on [0,1]?  I'd say
> the latter is way more usual when training networks as typically
> binary crossentropy yields better result, so that's what I'm using
> in https://github.com/pasky/michi/tree/nnet for the time being, but
> maybe I'm missing some good reason to use MSE instead?

Not that I know of. You can certainly get some networks to converge
better by using cross-entropy over MSE.

Maybe it's related to the nature of the errors? More avoidance of the
output being entirely wrong? Or habit? MSE is generally preferred for
regression-like problems, but you can argue whether a go position is
being regressed to some winrate%, or to win/loss...

-- 
GCP
___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] AlphaGo Zero self-play temperature

[Gian-Carlo Pascutto]

On 7/11/2017 19:07, Imran Hendley wrote:
> Am I understanding this correctly?

Yes.

It's possible they had in-betweens or experimented with variations at
some point, then settled on the simplest case. You can vary the
randomness if you define it as a softmax with varying temperature,
that's harder if you only define the policy as select best or select
proportionally.

-- 
GCP
___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] AlphaGo Zero self-play temperature

[uurtamo .]

If I understand your question correctly, "goes to 1" can happen as quickly
or slowly as you'd like. Yes?

On Nov 7, 2017 7:26 PM, "Imran Hendley"  wrote:

Hi, I might be having trouble understanding the self-play policy for
AlphaGo Zero. Can someone let me know if I'm on the right track here?

The paper states:

In each position s, an MCTS search is executed, guided by the neural
network f_θ . The
MCTS search outputs probabilities π of playing each move.


This wasn't clear at first since MCTS outputs wins and visits, but later
the paper explains further:

MCTS may be viewed as a self-play algorithm that, given neural
network parameters θ and a root position s, computes a vector of search
probabilities recommending moves to play, π =​  α_θ(s), proportional to
the exponentiated visit count for each move, π_a ∝​  N(s, a)^(1/τ) , where
τ is
a temperature parameter.


So this makes sense, but when I looked for the schedule for decaying the
temperature all I found was the following in the Self-play section of
Methods:


For the first 30 moves of each game, the temperature is set to τ = ​1; this
selects moves proportionally to their visit count in MCTS, and ensures a
diverse
set of positions are encountered. For the remainder of the game, an
infinitesimal
temperature is used, τ→​0.

This sounds like they are sampling proportional to visits for the first 30
moves since τ = ​1 makes the exponent go away, and after that they are
playing the move with the most visits, since the probability of the move
with the most visits goes to 1 and the probability of all other moves goes
to zero in the expression π(a | s_0) = N(s_0 , a)^(1/τ) / ∑ b N(s_0 ,
b)^(1/τ) as τ goes to 0 from the right.

Am I understanding this correctly? I am confused because it seems a little
convoluted to define this simple policy in terms of a temperature. When
they mentioned temperature I was expecting something that slowly decays
over time rather than only taking two trivial values.

Thanks!


___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go
___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] AlphaGo Zero self-play temperature

[Álvaro Begué]

Your understanding matches mine. My guess is that they had a temperature
parameter in the code that would allow for things like slowly transitioning
from random sampling to deterministically picking the maximum, but they
ended up using only those particular values.

Álvaro.




On Tue, Nov 7, 2017 at 1:07 PM, Imran Hendley 
wrote:

> Hi, I might be having trouble understanding the self-play policy for
> AlphaGo Zero. Can someone let me know if I'm on the right track here?
>
> The paper states:
>
> In each position s, an MCTS search is executed, guided by the neural
> network f_θ . The
> MCTS search outputs probabilities π of playing each move.
>
>
> This wasn't clear at first since MCTS outputs wins and visits, but later
> the paper explains further:
>
> MCTS may be viewed as a self-play algorithm that, given neural
> network parameters θ and a root position s, computes a vector of search
> probabilities recommending moves to play, π =​  α_θ(s), proportional to
> the exponentiated visit count for each move, π_a ∝​  N(s, a)^(1/τ) , where
> τ is
> a temperature parameter.
>
>
> So this makes sense, but when I looked for the schedule for decaying the
> temperature all I found was the following in the Self-play section of
> Methods:
>
>
> For the first 30 moves of each game, the temperature is set to τ = ​1; this
> selects moves proportionally to their visit count in MCTS, and ensures a
> diverse
> set of positions are encountered. For the remainder of the game, an
> infinitesimal
> temperature is used, τ→​0.
>
> This sounds like they are sampling proportional to visits for the first 30
> moves since τ = ​1 makes the exponent go away, and after that they are
> playing the move with the most visits, since the probability of the move
> with the most visits goes to 1 and the probability of all other moves goes
> to zero in the expression π(a | s_0) = N(s_0 , a)^(1/τ) / ∑ b N(s_0 ,
> b)^(1/τ) as τ goes to 0 from the right.
>
> Am I understanding this correctly? I am confused because it seems a little
> convoluted to define this simple policy in terms of a temperature. When
> they mentioned temperature I was expecting something that slowly decays
> over time rather than only taking two trivial values.
>
> Thanks!
>
>
> ___
> Computer-go mailing list
> Computer-go@computer-go.org
> http://computer-go.org/mailman/listinfo/computer-go
>
___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] AlphaGo Zero SGF - Free Use or Copyright?

[Pierce T. Wetter III]

Pardon my cynicism but I think Germany just guaranteed that other countries 
will develop self driving cars first and Germany will end up adapting someone 
elses solution after they’ve test driven it on _their_ citizens. Which may be 
their intent...

All of the self-driving car “knowledge" will be fuzzy. At best this rule makes 
lawyers rich.

On Oct 30, 2017, 11:36 PM -0700, Robert Jasiek , wrote:
> On 30.10.2017 19:22, Pierce T. Wetter III wrote:
> > this car and this child
>
> In Germany, an ethics commission has written ethical guidelines for
> self-driving cars with also the rule to always prefer avoiding
> casualties of human beings.
>
> --
> robert jasiek
> ___
> Computer-go mailing list
> Computer-go@computer-go.org
> http://computer-go.org/mailman/listinfo/computer-go
___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] AlphaGo Zero SGF - Free Use or Copyright?

[Thomas Rohde]

On 2017-11-01 at 11:48, adrian.b.rob...@gmail.com wrote:

> Robert Jasiek  writes:
> 
>>> [..] 
>> 
>> In Germany, an ethics commission has written ethical guidelines for 
>> self-driving cars with also the rule to always prefer avoiding casualties of 
>> human beings.
> 
> Did they consult Isaac Asimov on this?

Should actually have been their first thought, IMO (and I guess they were), 
Asimov’s Laws of Robotics should be obligatory reading for everybody in 
robotics. (and probably should be expanded/modernised — adapted to contemporary 
knowledge)


> Jokes aside, it's interesting to see some convergence of science to science 
> fiction..

Not surprising, IMO, since many of the best SF writers, including Asimov, 
actually were and are scientists and/or engineers. And scientists (as well as 
engineers, even industrial designers) often take their ideas from SF, again 
unsurprising … I’d assume that reading SF has inspired many people to pursue 
science, engineering, astronautics, etc.

Also, there is the word “science” in Science Fiction, which should tell us 
something. SF is not necessarily always “space opera” with goodies and baddies 
and alien monsters, space opera is often not much more than cowboy stories in 
“spacey” apparel.

Forgive the off-topic … as an avid SF reader I was triggered ;-)


Respectfully,
Tom
___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] AlphaGo Zero SGF - Free Use or Copyright?

[Adrian . B . Robert]

Robert Jasiek  writes:

> On 30.10.2017 19:22, Pierce T. Wetter III wrote:
>> this car and this child
>
> In Germany, an ethics commission has written ethical guidelines for
> self-driving cars with also the rule to always prefer avoiding
> casualties of human beings.

Did they consult Isaac Asimov on this?

Jokes aside, it's interesting to see some convergence of science
to science fiction..

___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] AlphaGo Zero SGF - Free Use or Copyright?

[Marc Landgraf]

There is even a decent site for those situations:

http://moralmachine.mit.edu/ (select language and then click "start
judging")

2017-10-31 7:55 GMT+01:00 Petri Pitkanen :

> and we can allways come up with bizarre situation like casualties insidet
> the vehicle vs casualties to persoons outside the vehicle. I am pretty sure
> this will a long discussion with huge research gaps on ethics  as well as
> in engineering
>
> 2017-10-31 7:00 GMT+02:00 Robert Jasiek :
>
>> On 30.10.2017 19:22, Pierce T. Wetter III wrote:
>>
>>> this car and this child
>>>
>>
>> In Germany, an ethics commission has written ethical guidelines for
>> self-driving cars with also the rule to always prefer avoiding casualties
>> of human beings.
>>
>>
>> --
>> robert jasiek
>> ___
>> Computer-go mailing list
>> Computer-go@computer-go.org
>> http://computer-go.org/mailman/listinfo/computer-go
>>
>
>
> ___
> Computer-go mailing list
> Computer-go@computer-go.org
> http://computer-go.org/mailman/listinfo/computer-go
>
___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] AlphaGo Zero SGF - Free Use or Copyright?

[Lucas Baker]

Actually, you can try impersonating the AI yourself:
http://moralmachine.mit.edu/

On Tue, Oct 31, 2017 at 5:16 PM Petri Pitkanen 
wrote:

> and we can allways come up with bizarre situation like casualties insidet
> the vehicle vs casualties to persoons outside the vehicle. I am pretty sure
> this will a long discussion with huge research gaps on ethics  as well as
> in engineering
>
> 2017-10-31 7:00 GMT+02:00 Robert Jasiek :
>
>> On 30.10.2017 19:22, Pierce T. Wetter III wrote:
>>
>>> this car and this child
>>>
>>
>> In Germany, an ethics commission has written ethical guidelines for
>> self-driving cars with also the rule to always prefer avoiding casualties
>> of human beings.
>>
>>
>> --
>> robert jasiek
>> ___
>> Computer-go mailing list
>> Computer-go@computer-go.org
>> http://computer-go.org/mailman/listinfo/computer-go
>>
>
> ___
> Computer-go mailing list
> Computer-go@computer-go.org
> http://computer-go.org/mailman/listinfo/computer-go
___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] AlphaGo Zero SGF - Free Use or Copyright?

[Petri Pitkanen]

and we can allways come up with bizarre situation like casualties insidet
the vehicle vs casualties to persoons outside the vehicle. I am pretty sure
this will a long discussion with huge research gaps on ethics  as well as
in engineering

2017-10-31 7:00 GMT+02:00 Robert Jasiek :

> On 30.10.2017 19:22, Pierce T. Wetter III wrote:
>
>> this car and this child
>>
>
> In Germany, an ethics commission has written ethical guidelines for
> self-driving cars with also the rule to always prefer avoiding casualties
> of human beings.
>
>
> --
> robert jasiek
> ___
> Computer-go mailing list
> Computer-go@computer-go.org
> http://computer-go.org/mailman/listinfo/computer-go
>
___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] AlphaGo Zero SGF - Free Use or Copyright?

[Robert Jasiek]

On 30.10.2017 19:22, Pierce T. Wetter III wrote:

this car and this child


In Germany, an ethics commission has written ethical guidelines for 
self-driving cars with also the rule to always prefer avoiding 
casualties of human beings.


--
robert jasiek
___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] AlphaGo Zero SGF - Free Use or Copyright?

[Álvaro Begué]

I am not sure how people are designing self-driving cars, but if it were up
to me, it would be very explicitly about maximizing expected utility. A
neural network can be trained to estimate the expected sum of future
rewards, usually with some exponential future discount. Actually, that's
explicitly what Q-learning does, and it's not that different from how
AlphaGo's value network works.

The fact that it's hard to figure out why a neural network did what it did
is not worse than the situation with humans. We don't understand neurology
well enough to know why someone didn't see a pedestrian or a red light. And
somehow the legal system doesn't collapse. In the case of neural networks,
the case that resulted in the accident and similar cases can be added to
the training database to make future versions of the network more robust,
so over time the number of accidents should drop fast.

Álvaro.



On Mon, Oct 30, 2017 at 6:06 PM, Pierce T. Wetter III <
pie...@alumni.caltech.edu> wrote:

> I would argue that if I was an engineer for a hypothetical autonomous car
> manufacturer, that it would be critically important to keep a running
> circular buffer of all the inputs over time for the car. Sort of like how
> existing cars have Dash Cams that continuously record to flash, but only
> keep the video if you tell it to or it detects major G forces.
>
> To your point, I’m not sure the car would necessarily be able to tell tree
> from child, tree might be “certain large obstacle” and child is “smaller
> large obstacle”. So that would give them the same utility function -1000.
> But utility functions are rarely so straightforward in a neural network as
> you suppose.
>
> I think it would take differential analysis (A term I just made up) to
> determine the utility function, which is why having a continuous log of all
> the input streams is necessary.
>
> On Oct 30, 2017, 3:45 PM -0700, Álvaro Begué ,
> wrote:
>
> In your hypothetical scenario, if the car can give you as much debugging
> information as you suggest (100% tree is there, 95% child is there), you
> can actually figure out what's happening. The only other piece of
> information you need is the configured utility values for the possible
> outcomes.
>
> Say the utility of hitting a tree is -1000, the utility of hitting a child
> is -5000 and the utility of not hitting anything is 0. A rational agent
> maximizes the expected value of the utility function. So:
>  - Option A: Hit the tree. Expected utility = -1000.
>  - Option B: Avoid the tree, possibly hitting the child, if there is a
> child there after all. Expected utility: 0.95 * (-5000) + 0.05 * 0 = -4750.
>
> So the car should pick option A. If the configured utility function is
> such that hitting a tree and hitting a child have the same value, the
> lawyers would be correct that the programmers are endangering the public
> with their bad programming.
>
> Álvaro.
>
>
>
> On Mon, Oct 30, 2017 at 2:22 PM, Pierce T. Wetter III <
> pie...@alumni.caltech.edu> wrote:
>
>> Unlike humans, who have these pesky things called rights, we can abuse
>> our computer programs to deduce why they made decisions. I can see a future
>> where that has to happen. From my experience in trying to best the stock
>> market with an algorithm I can tell you that you have to be able to explain
>> why something happened, or the CEO will rest control away from the
>> engineers.
>>
>> Picture a court case where the engineers for an electric car are called
>> upon to testify about why a child was killed by their self driving car. The
>> fact that the introduction of the self-driving car has reduced the accident
>> rate by 99% doesn’t matter, because the court case is about *this* car
>> and *this* child. The 99% argument is for the closing case, or for the
>> legislature, but it’s early yet.
>>
>> The Manufacturer throws up their arms and says “we dunno, sorry”.
>>
>> Meanwhile, the plaintiff has hired someone who has manipulated the inputs
>> to the neural net, and they’ve figured out that the car struck the child,
>> because the car was 100% sure the tree was there, but it could only be 95%
>> sure the child was there. So it ruthlessly aimed for the lesser
>> probability.
>>
>> The plaintiff’s lawyer argues that a human would have rather hit a tree
>> than a child.
>>
>> Jury awards $100M in damages to the plaintiffs.
>>
>> I would think it would be possible to do “differential” analysis on AGZ
>> positions to see why AGZ made certain moves. Add an eye to a weak group,
>> etc. Essentially that’s what we’re doing with MCTS, right?
>>
>> It seems like a fun research project to try to build a system that can
>> reverse engineer AGZ, and not only would it be fun, but its a moral
>> imperative.
>>
>> Pierce
>>
>>
>> ___
>> Computer-go mailing list
>> Computer-go@computer-go.org
>> http://computer-go.org/mailman/listinfo/computer-go
>>
>
> 

Re: [Computer-go] AlphaGo Zero SGF - Free Use or Copyright?

[Pierce T. Wetter III]

I would argue that if I was an engineer for a hypothetical autonomous car 
manufacturer, that it would be critically important to keep a running circular 
buffer of all the inputs over time for the car. Sort of like how existing cars 
have Dash Cams that continuously record to flash, but only keep the video if 
you tell it to or it detects major G forces.

To your point, I’m not sure the car would necessarily be able to tell tree from 
child, tree might be “certain large obstacle” and child is “smaller large 
obstacle”. So that would give them the same utility function -1000. But utility 
functions are rarely so straightforward in a neural network as you suppose.

I think it would take differential analysis (A term I just made up) to 
determine the utility function, which is why having a continuous log of all the 
input streams is necessary.

On Oct 30, 2017, 3:45 PM -0700, Álvaro Begué , wrote:
> In your hypothetical scenario, if the car can give you as much debugging 
> information as you suggest (100% tree is there, 95% child is there), you can 
> actually figure out what's happening. The only other piece of information you 
> need is the configured utility values for the possible outcomes.
>
> Say the utility of hitting a tree is -1000, the utility of hitting a child is 
> -5000 and the utility of not hitting anything is 0. A rational agent 
> maximizes the expected value of the utility function. So:
>  - Option A: Hit the tree. Expected utility = -1000.
>  - Option B: Avoid the tree, possibly hitting the child, if there is a child 
> there after all. Expected utility: 0.95 * (-5000) + 0.05 * 0 = -4750.
>
> So the car should pick option A. If the configured utility function is such 
> that hitting a tree and hitting a child have the same value, the lawyers 
> would be correct that the programmers are endangering the public with their 
> bad programming.
>
> Álvaro.
>
>
>
> > On Mon, Oct 30, 2017 at 2:22 PM, Pierce T. Wetter III 
> >  wrote:
> > > Unlike humans, who have these pesky things called rights, we can abuse 
> > > our computer programs to deduce why they made decisions. I can see a 
> > > future where that has to happen. From my experience in trying to best the 
> > > stock market with an algorithm I can tell you that you have to be able to 
> > > explain why something happened, or the CEO will rest control away from 
> > > the engineers.
> > >
> > > Picture a court case where the engineers for an electric car are called 
> > > upon to testify about why a child was killed by their self driving car. 
> > > The fact that the introduction of the self-driving car has reduced the 
> > > accident rate by 99% doesn’t matter, because the court case is about this 
> > > car and this child. The 99% argument is for the closing case, or for the 
> > > legislature, but it’s early yet.
> > >
> > > The Manufacturer throws up their arms and says “we dunno, sorry”.
> > >
> > > Meanwhile, the plaintiff has hired someone who has manipulated the inputs 
> > > to the neural net, and they’ve figured out that the car struck the child, 
> > > because the car was 100% sure the tree was there, but it could only be 
> > > 95% sure the child was there. So it ruthlessly aimed for the lesser 
> > > probability.
> > >
> > > The plaintiff’s lawyer argues that a human would have rather hit a tree 
> > > than a child.
> > >
> > > Jury awards $100M in damages to the plaintiffs.
> > >
> > > I would think it would be possible to do “differential” analysis on AGZ 
> > > positions to see why AGZ made certain moves. Add an eye to a weak group, 
> > > etc. Essentially that’s what we’re doing with MCTS, right?
> > >
> > > It seems like a fun research project to try to build a system that can 
> > > reverse engineer AGZ, and not only would it be fun, but its a moral 
> > > imperative.
> > >
> > > Pierce
> > >
> > >
> > > ___
> > > Computer-go mailing list
> > > Computer-go@computer-go.org
> > > http://computer-go.org/mailman/listinfo/computer-go
>
> ___
> Computer-go mailing list
> Computer-go@computer-go.org
> http://computer-go.org/mailman/listinfo/computer-go
___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] AlphaGo Zero SGF - Free Use or Copyright?

[Álvaro Begué]

In your hypothetical scenario, if the car can give you as much debugging
information as you suggest (100% tree is there, 95% child is there), you
can actually figure out what's happening. The only other piece of
information you need is the configured utility values for the possible
outcomes.

Say the utility of hitting a tree is -1000, the utility of hitting a child
is -5000 and the utility of not hitting anything is 0. A rational agent
maximizes the expected value of the utility function. So:
 - Option A: Hit the tree. Expected utility = -1000.
 - Option B: Avoid the tree, possibly hitting the child, if there is a
child there after all. Expected utility: 0.95 * (-5000) + 0.05 * 0 = -4750.

So the car should pick option A. If the configured utility function is such
that hitting a tree and hitting a child have the same value, the lawyers
would be correct that the programmers are endangering the public with their
bad programming.

Álvaro.



On Mon, Oct 30, 2017 at 2:22 PM, Pierce T. Wetter III <
pie...@alumni.caltech.edu> wrote:

> Unlike humans, who have these pesky things called rights, we can abuse our
> computer programs to deduce why they made decisions. I can see a future
> where that has to happen. From my experience in trying to best the stock
> market with an algorithm I can tell you that you have to be able to explain
> why something happened, or the CEO will rest control away from the
> engineers.
>
> Picture a court case where the engineers for an electric car are called
> upon to testify about why a child was killed by their self driving car. The
> fact that the introduction of the self-driving car has reduced the accident
> rate by 99% doesn’t matter, because the court case is about *this* car
> and *this* child. The 99% argument is for the closing case, or for the
> legislature, but it’s early yet.
>
> The Manufacturer throws up their arms and says “we dunno, sorry”.
>
> Meanwhile, the plaintiff has hired someone who has manipulated the inputs
> to the neural net, and they’ve figured out that the car struck the child,
> because the car was 100% sure the tree was there, but it could only be 95%
> sure the child was there. So it ruthlessly aimed for the lesser
> probability.
>
> The plaintiff’s lawyer argues that a human would have rather hit a tree
> than a child.
>
> Jury awards $100M in damages to the plaintiffs.
>
> I would think it would be possible to do “differential” analysis on AGZ
> positions to see why AGZ made certain moves. Add an eye to a weak group,
> etc. Essentially that’s what we’re doing with MCTS, right?
>
> It seems like a fun research project to try to build a system that can
> reverse engineer AGZ, and not only would it be fun, but its a moral
> imperative.
>
> Pierce
>
>
> ___
> Computer-go mailing list
> Computer-go@computer-go.org
> http://computer-go.org/mailman/listinfo/computer-go
>
___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] AlphaGo Zero SGF - Free Use or Copyright?

[Pierce T. Wetter III]

Unlike humans, who have these pesky things called rights, we can abuse our 
computer programs to deduce why they made decisions. I can see a future where 
that has to happen. From my experience in trying to best the stock market with 
an algorithm I can tell you that you have to be able to explain why something 
happened, or the CEO will rest control away from the engineers.

Picture a court case where the engineers for an electric car are called upon to 
testify about why a child was killed by their self driving car. The fact that 
the introduction of the self-driving car has reduced the accident rate by 99% 
doesn’t matter, because the court case is about this car and this child. The 
99% argument is for the closing case, or for the legislature, but it’s early 
yet.

The Manufacturer throws up their arms and says “we dunno, sorry”.

Meanwhile, the plaintiff has hired someone who has manipulated the inputs to 
the neural net, and they’ve figured out that the car struck the child, because 
the car was 100% sure the tree was there, but it could only be 95% sure the 
child was there. So it ruthlessly aimed for the lesser probability.

The plaintiff’s lawyer argues that a human would have rather hit a tree than a 
child.

Jury awards $100M in damages to the plaintiffs.

I would think it would be possible to do “differential” analysis on AGZ 
positions to see why AGZ made certain moves. Add an eye to a weak group, etc. 
Essentially that’s what we’re doing with MCTS, right?

It seems like a fun research project to try to build a system that can reverse 
engineer AGZ, and not only would it be fun, but its a moral imperative.

Pierce

___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] AlphaGo Zero SGF - Free Use or Copyright?

[terry mcintyre via Computer-go]

Petri and Tom are correct; "intuition" and "subconscious" and "unobserved 
thought" are names for the same idea. 
AlphaGo Zero's neural network, regardless of how well it simulates human 
neurons or not, cannot be described as being similar to what we think of as the 
logical, rule-driven part of the human thought process; it is much more akin to 
the "intuition" or "subconscious" of a highly experienced player, wrapped 
together with a very logical process akin to what we humans call "reading." 
Terry McIntyre  Unix/Linux Systems Administration 
Taking time to do it right saves having to do it twice. 

On Sunday, October 29, 2017, 6:42:27 AM EDT, Petri Pitkanen 
 wrote:  
 
 intuition is handy word for truly automated information processing i.e 
subconscious.   And everything that train conscious decission making trains 
also the subconscious/intuiton. Intuiton nothing mythical just automation 
achieved via training
2017-10-29 5:08 GMT+02:00 Thomas Rohde :

On 2017-10-28 at 16:36, Robert Jasiek  wrote:

> IMO, intuition does not exist; it is nothing but an excuse for not 
> understanding subconscious or currently unobservable thinking yet. Can we 
> speak of human subconscious thinking, please?

Uhm, I always thought the short word for “subconscious thinking” was 
“intuition” ;-)

Reminds me of “A Table is a Table” (orig. “Ein Tisch ist ein Tisch”), a short 
story by Swiss writer Peter Bichsel

—> https://vimeo.com/11331609 (ten minutes video, English version)
—> https://vimeo.com/8749843 (German version)

“What's in a name? that which we call a rose
By any other word would smell as sweet”
— Shakespeare


Greetings, Tom

--
Thomas Rohde
Wiesenkamp 12, 29646 Bispingen, GERMANY
--
t...@bonobo.com
__ _
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/ mailman/listinfo/computer-go

___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] AlphaGo Zero SGF - Free Use or Copyright?

[Petri Pitkanen]

intuition is handy word for truly automated information processing i.e
subconscious.   And everything that train conscious decission making trains
also the subconscious/intuiton. Intuiton nothing mythical just automation
achieved via training

2017-10-29 5:08 GMT+02:00 Thomas Rohde :

> On 2017-10-28 at 16:36, Robert Jasiek  wrote:
>
> > IMO, intuition does not exist; it is nothing but an excuse for not
> understanding subconscious or currently unobservable thinking yet. Can we
> speak of human subconscious thinking, please?
>
> Uhm, I always thought the short word for “subconscious thinking” was
> “intuition” ;-)
>
> Reminds me of “A Table is a Table” (orig. “Ein Tisch ist ein Tisch”), a
> short story by Swiss writer Peter Bichsel
>
> —> https://vimeo.com/11331609 (ten minutes video, English version)
> —> https://vimeo.com/8749843 (German version)
>
> “What's in a name? that which we call a rose
> By any other word would smell as sweet”
> — Shakespeare
>
>
> Greetings, Tom
>
> --
> Thomas Rohde
> Wiesenkamp 12, 29646 Bispingen, GERMANY
> --
> t...@bonobo.com
> ___
> Computer-go mailing list
> Computer-go@computer-go.org
> http://computer-go.org/mailman/listinfo/computer-go
>
___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] AlphaGo Zero SGF - Free Use or Copyright?

[Thomas Rohde]

On 2017-10-28 at 16:36, Robert Jasiek  wrote:

> IMO, intuition does not exist; it is nothing but an excuse for not 
> understanding subconscious or currently unobservable thinking yet. Can we 
> speak of human subconscious thinking, please?

Uhm, I always thought the short word for “subconscious thinking” was 
“intuition” ;-)

Reminds me of “A Table is a Table” (orig. “Ein Tisch ist ein Tisch”), a short 
story by Swiss writer Peter Bichsel

—> https://vimeo.com/11331609 (ten minutes video, English version)
—> https://vimeo.com/8749843 (German version)

“What's in a name? that which we call a rose
By any other word would smell as sweet”
— Shakespeare


Greetings, Tom

-- 
Thomas Rohde
Wiesenkamp 12, 29646 Bispingen, GERMANY
--
t...@bonobo.com
___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] AlphaGo Zero SGF - Free Use or Copyright?

[Robert Jasiek]

On 28.10.2017 11:13, Petri Pitkanen wrote:

Exactly verbalized rules lose to pure analysis power.


(I think with "verbalised" you mean "codified in writing", with "pure 
analysis power" you mean "volume of reading, calculation, sampling or NN 
processing".)


Rules are not meant to win or lose against "pure analysis power" but to 
use it when necessary and unavoidable, e.g., tactical reading when 
clarifying L+D status. A rule can be "Consider an attack if the L+D 
status is 'unsettled'" but also tactical reading determines that status.



Human intuition is trained with endless repetition.


IMO, intuition does not exist; it is nothing but an excuse for not 
understanding subconscious or currently unobservable thinking yet. Can 
we speak of human subconscious thinking, please?


Subconscious thinking can be trained by learning rules, practising 
problems etc. Conscious, explicit thinking can be trained by learning 
rules, practising problems etc. So what do you want to say?


--
robert jasiek
___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] AlphaGo Zero SGF - Free Use or Copyright?

[Petri Pitkanen]

Exactly verbalized rules lose to pure analysis power. Though much chess
intuiton is coded into  evaluation function. Buiding analysis trees to
alfa-beta pruning BUT in quite differently human woudl do it, just basic
idea/ideas are there.

Human intuition is trained with endless repetition. Like IM Jeremy Silman
who went through about hundred games a night while teenager (quite a feat
on actual board) to 'train' his pattern matcher.

I do doubt if anyone coudl codifly that information in fully transferrable
way at all. In teacher-pupil interaction somehow. But as an book, noway.
Hard to say what chess IM is go terms but whole chess Grand master to
Master ranks are within 300 elopoints-. And in upper echelons one Dan rank
is about 250-300 elopoints so IM woudl strongish 6dan perhaps , not quite
7dan.  So 4dan is way better than what I can drema of but still is chess
ranks that woudl be a like Elo 2000 a good player but no way near a master.
So I woudl say that the old way, how ever tedious and non-analytical is
still required to reach the top of game.

But then again teaching method to quickly reach a reasonable strength is
certainly needed. Mayre robert has it, do not know as have not tried

2017-10-28 1:39 GMT+03:00 uurtamo . :

> By way of comparison.
>
> It would be ludicrous to ask a world champion chess player to explain
> their strategy in a "programmable" way. it would certainly result in a
> player much worse than the best computer player, if it were to be coded up,
> even if you spent 40 years decoding intuition, etc, and got it exactly
> correct.
>
> Why do I say this? Because the best human player will lose > 90% of the
> time against the best computer player. And they understand their own
> intuition fairly well.
>
> Do we want to sit down and analyze the best human player's intuition?
> Perhaps. But certainly not to improve the best computer player. It can
> already crush all humans at pretty much every strength.
>
> s.
>
>
> On Fri, Oct 27, 2017 at 10:37 AM, Robert Jasiek  wrote:
>
>> On 27.10.2017 13:58, Petri Pitkanen wrote:
>>
>>> doubt that your theory is any better than some competing ones.
>>>
>>
>> For some specialised topics, it is evident that my theory is better or
>> belongs to the few applicable theories (often by other amateur-player
>> researchers) worth considering.
>>
>> For a broad sense of "covering every aspect of go theory", I ask: what
>> competing theories? E.g., take verbal theory teaching by professional
>> players and they say, e.g., "Follow the natural flow of the game". I have
>> heard this for decades but still do not have the slightest idea what it
>> might mean. It assumes meaning only if I replace it by my theory. Or they
>> say: "Respect the beauty of shapes!" I have no idea what this means.
>>
>> A few particular professional players have reasonable theories on
>> specific topics and resembling methodical approach occurring in my theories.
>>
>> So what competing theories do you mean?
>>
>> The heritage of professional shape examples? If you want to call that
>> theory.
>>
>> As I do know people who are stronger than you and are using different
>>> framework.
>>>
>>
>> Yes, but where do they describe it? Almost all professional players I
>> have asked to explain their decision-making have said that they could not
>> because it would be intuition. A framework that is NOT theory.
>>
>>
>> --
>> robert jasiek
>> ___
>> Computer-go mailing list
>> Computer-go@computer-go.org
>> http://computer-go.org/mailman/listinfo/computer-go
>>
>
>
> ___
> Computer-go mailing list
> Computer-go@computer-go.org
> http://computer-go.org/mailman/listinfo/computer-go
>
___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] AlphaGo Zero SGF - Free Use or Copyright?

[uurtamo .]

By way of comparison.

It would be ludicrous to ask a world champion chess player to explain their
strategy in a "programmable" way. it would certainly result in a player
much worse than the best computer player, if it were to be coded up, even
if you spent 40 years decoding intuition, etc, and got it exactly correct.

Why do I say this? Because the best human player will lose > 90% of the
time against the best computer player. And they understand their own
intuition fairly well.

Do we want to sit down and analyze the best human player's intuition?
Perhaps. But certainly not to improve the best computer player. It can
already crush all humans at pretty much every strength.

s.


On Fri, Oct 27, 2017 at 10:37 AM, Robert Jasiek  wrote:

> On 27.10.2017 13:58, Petri Pitkanen wrote:
>
>> doubt that your theory is any better than some competing ones.
>>
>
> For some specialised topics, it is evident that my theory is better or
> belongs to the few applicable theories (often by other amateur-player
> researchers) worth considering.
>
> For a broad sense of "covering every aspect of go theory", I ask: what
> competing theories? E.g., take verbal theory teaching by professional
> players and they say, e.g., "Follow the natural flow of the game". I have
> heard this for decades but still do not have the slightest idea what it
> might mean. It assumes meaning only if I replace it by my theory. Or they
> say: "Respect the beauty of shapes!" I have no idea what this means.
>
> A few particular professional players have reasonable theories on specific
> topics and resembling methodical approach occurring in my theories.
>
> So what competing theories do you mean?
>
> The heritage of professional shape examples? If you want to call that
> theory.
>
> As I do know people who are stronger than you and are using different
>> framework.
>>
>
> Yes, but where do they describe it? Almost all professional players I have
> asked to explain their decision-making have said that they could not
> because it would be intuition. A framework that is NOT theory.
>
>
> --
> robert jasiek
> ___
> Computer-go mailing list
> Computer-go@computer-go.org
> http://computer-go.org/mailman/listinfo/computer-go
>
___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] AlphaGo Zero SGF - Free Use or Copyright?

[Robert Jasiek]

On 27.10.2017 13:58, Petri Pitkanen wrote:

doubt that your theory is any better than some competing ones.


For some specialised topics, it is evident that my theory is better or 
belongs to the few applicable theories (often by other amateur-player 
researchers) worth considering.


For a broad sense of "covering every aspect of go theory", I ask: what 
competing theories? E.g., take verbal theory teaching by professional 
players and they say, e.g., "Follow the natural flow of the game". I 
have heard this for decades but still do not have the slightest idea 
what it might mean. It assumes meaning only if I replace it by my 
theory. Or they say: "Respect the beauty of shapes!" I have no idea what 
this means.


A few particular professional players have reasonable theories on 
specific topics and resembling methodical approach occurring in my theories.


So what competing theories do you mean?

The heritage of professional shape examples? If you want to call that 
theory.



As I do know people who are stronger than you and are using different
framework.


Yes, but where do they describe it? Almost all professional players I 
have asked to explain their decision-making have said that they could 
not because it would be intuition. A framework that is NOT theory.


--
robert jasiek
___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] AlphaGo Zero SGF - Free Use or Copyright?

[Petri Pitkanen]

You playing strength is anecdotal evidence. And quite often going through
just systematic way your thinking is more valuable than the actual end
product. As it programs you subconscious decision making. You said that it
is not part of your decision making but that is unlikely to be true. People
do not know when subconscious decision are made as the upper layer
rationalizes the decisions afterwards.
https://www.relationshipscoach.co.uk/blog/research-shows-our-subconscious-mind-makes-our-decisions-for-us/

and that is not bad. Your huge effort to become strong player did program
you intuitive decision making to such degree that it is worth listening.

I still would doubt that your theory is any better than some competing
ones. As I do know people who are stronger than you and are using different
framework. Similarity is the directed and intentional search of truth.
Process is probably way more important the result. Obviousl I canno tprove
my point as my evidence is anecdotal

PP

2017-10-26 17:54 GMT+03:00 Robert Jasiek :

> On 26.10.2017 08:52, Petri Pitkanen wrote:
>
>> Unfortunately there is no proof that you principles work better than those
>> form eighties.
>>
>
> No computer-go proof.
>
> There is evidence in the form of my playing strength: with the principles
> "from the eighties", I got to circa 1 kyu. L+D reading practice etc. made
> me 3 dan. Afterwards, almost the only thing that made me stronger to 5 dan
> and then further improved my understanding was the invention of my own
> principles.
>
> My principles etc. also work for (an unknown fraction of) readers of my
> books and for a high percentage of my pupils but I cannot compare what the
> effect on them would have been if instead they would only have learnt the
> principles "from the eighties". I do, however, know that my principles
> provide me with very much more efficient means of teaching contents
> compared to using the principles "from the eighties".
>
> The principles "from the eighties" and my principles can be compared with
> each other. IMO, such a comparison is shocking: the principles "from the
> eighties" are very much weaker on average and altogether convey very much
> less contents.
>
> Nor there is any agreement that your pronciples form any
>> improvement over the old ones.
>>
>
> Only time constraints prevent me from doing an extensive comparison and so
> better support formation of an agreement.
>
> What is missing that I doubt that you can verbalise your go understanding
>> to degree that by applying those principles  I could become substantially
>> better player.
>>
>
> Different players are different. So different that some players claim to
> only learn from examples. Therefore, I cannot know whether you are a player
> who could learn well from principles etc.
>
> - My reading skills would not get any better
>>
>
> Do you say so after having learnt and invested effort in applying the
> contents of Tactical Reading?
>
> Regardless of the possible impact of that book, a great part of reading
> skill must be obtained by reading practice in games and problem solving. If
> your reading is much weaker than your knowledge of go theory, then it may
> be the case that almost only reading practise (plus possibly reading theory
> about improving one's reading practice) can significantly improve your
> strength at the moment.
>
> - your principles are more complex than you understand.
>>
>
> I do not think so:)
>
> Much of you know is
>> automated to degree that it is subconsciousness information.
>>
>
> From ca. 10 kyu to now, especially from 3 dan to now, I have reduced the
> impact of my subconscious thinking on my go decision-making and replaced it
> by knowledge, reading and positional judgement based on knowledge and
> reading. The still remaining subconscious thinking is small. Most of my
> remaining mistakes are related to psychology or subconscious thinking, when
> necessary because of explicit knowledge gaps or thinking time constraints.
>
> Transferring that information if hard.
>>
>
> Transferring it from principles etc. to code - yes.
>
> If you can build Go bot about  KGS 3/4dan strength
>>
>
> Using my approach, I expect several manyears, which I do not have for that
> purpose.
>
>
> --
> robert jasiek
> ___
> Computer-go mailing list
> Computer-go@computer-go.org
> http://computer-go.org/mailman/listinfo/computer-go
>
___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] AlphaGo Zero SGF - Free Use or Copyright?

[Jim O'Flaherty]

It's related to this line of thinking by Douglas Hoffstadter:
https://en.wikipedia.org/wiki/Copycat_(software)


Namaste,

Jim O'Flaherty
Founder/CEO
Precision Location Intelligence, Inc.
 • Irving, TX, USA
469-358-0633 <4693580633> • jim.oflaherty...@gmail.com •
www.linkedin.com/in/jimoflahertyjr

CONFIDENTIALITY / PROPRIETARY NOTICE:
The information contained in this e-mail, including any attachment(s), is
confidential information for Precision Location Intelligence, Inc.. As
such, it may be privileged and exempt from disclosure under applicable law.
If the reader of this message is not the intended recipient, or if you
received this message in error, then any direct or indirect disclosure,
distribution or copying of this message is strictly prohibited. If you have
received this message in error, please notify Precision Location
Intelligence, Inc. by calling (214) 489-7779 <2144897779> immediately and
by sending a return e-mail; delete this message; and destroy all copies,
including attachments.

On Thu, Oct 26, 2017 at 11:43 AM, Xavier Combelle  wrote:

> what are semantic genetic algorithm ?
>
> to my knowledge genetic algorithm lead to poor result except as a
> metaheuristic in optimisation problem
>
> Le 26/10/2017 à 14:40, Jim O'Flaherty a écrit :
>
> When I get time to spend dozens of hours on computer go again, I plan to
> play in Robert's area with semantic genetic algorithms. I am an Architect
> Software Engineer. Robert's work will allow me better than starting
> entirely from random in much the same way AlphaGo bootstrapped from the
> 100K of professional games. AG0 then leveraged AlphaGo in knowing an
> architecture that was close enough. My intuition is my approach will be
> something similar in it's evolution.
>
> This is the way we're going to "automate" creating provided proofing of
> human cognition styled computer go players to assist humans in a gradient
> ascent learning cycle.
>
> So, Robert, I admire and am encouraged by your research for my own
> computer go projects in this area. Keep kicking butt in your unique way. We
> are in an interesting transition in this community. Stick it out. It will
> be worth it long term.
>
> On Oct 26, 2017 4:38 AM, "Petri Pitkanen" 
> wrote:
>
>> Unfortunately there is no proof that you principles work better than
>> those form eighties. Nor there is any agreement that your pronciples form
>> any improvement over the old ones. Yes you are a  far better player than me
>> and shows that you are
>> - way better at reading
>> - have hugely better go understanding, principles if you like
>>
>> What is missing that I doubt that you can verbalise your go understanding
>> to degree that by applying those principles  I could become substantially
>> better player. again bulleting
>> - My reading skills would not get any better hence making much of value
>> any learning moot. Obviously issue on me not on your principles
>> - your principles are more complex than you understand. Much of you know
>> is automated to degree that it is subconsciousness information.
>> Transferring that information if hard. Usually done by re-playing master
>> games looking at problems i.e. training the darn neural net in the head
>>
>> If you can build Go bot about  KGS 3/4dan strength I am more than willing
>> to admit you are right and would even consider buying your  books.
>>
>> Petri
>>
>> 2017-10-26 6:21 GMT+03:00 Robert Jasiek :
>>
>>> On 25.10.2017 18:17, Xavier Combelle wrote:
>>>
 exact go theory is full of hole.

>>>
>>> WRT describing the whole game, yes, this is the current state. Solving
>>> go in a mathematical sense is a project for centuries.
>>>
>>> Actually, to my knowledge human can't apply only the exact go theory and
 play a decent game.

>>>
>>> Only for certain positions of a) late endgame, b) semeais, c) ko.
>>>
>>> If human can't do that, how it will teach a computer to do it magically ?

>>>
>>> IIRC, Martin Müller implemented CGT endgames a la Mathematical Go
>>> Endgames.
>>>
>>> The reason why (b) had became unpopular is because there is no go theory
 precise enough to implement it as an algorithm

>>>
>>> There is quite some theory of the 95% principle kind which might be
>>> implemented as approximation. E.g. "Usually, defend your weak important
>>> group." can be approximated by approximating "group", "important" (its loss
>>> is too large in a quick positional judgement), "weak" (can be killed in two
>>> successive moves), "defend" (after the move, cannot be killed in two
>>> successive moves), "usually" (always, unless there are several such groups
>>> and some must be chosen, say, randomly; the approximation being that the
>>> alternative strategy of large scale exchange is discarded).
>>>
>>> Besides, one must prioritise principles to solve conflicting principles
>>> by a higher order principle.

Re: [Computer-go] AlphaGo Zero SGF - Free Use or Copyright?

[Brian Sheppard via Computer-go]

Well,... good luck with that! :-)

Seriously: it is important to account for p-space completeness. That is, a set 
of rules that covers Go without conflict must be exponential in space usage.

Search has a triple role in system design. It accounts (at least 
asymptotically) for missing knowledge and also papers over disagreements 
between rules. It also evaluates the global situation, which allows rules to be 
expressed in terms of purely local effects.

From my perspective, that is too good a deal to pass by. But I don't want to be 
only a bearer of bad news. If you accept a limitation on your rule sets, then 
there is a higher level conflict resolution method that will lead to good 
results.

Your rules could express their effect as a local point gain, in the sense of 
"temperature". That is, temperature == the difference between moving first and 
letting the opponent move first. Then CGT provides a higher-order theory for 
rationalizing multiple priorities.

This suggestion only addresses one of the three roles of search, though perhaps 
the most important one.

Best,
Brian


-Original Message-
From: Computer-go [mailto:computer-go-boun...@computer-go.org] On Behalf Of 
Robert Jasiek
Sent: Thursday, October 26, 2017 10:17 AM
To: computer-go@computer-go.org
Subject: Re: [Computer-go] AlphaGo Zero SGF - Free Use or Copyright?

On 26.10.2017 13:52, Brian Sheppard via Computer-go wrote:
> MCTS is the glue that binds incompatible rules.

This is, however, not what I mean. Conflicting principles (call them rules if 
you like) must be dissolved by higher order principles. Only when all conflicts 
are dissolved should MCTS be applied.

What you describe has been used with success and better success than I expect 
what my knowledge-pure approach can currently achieve. But MCTS as glue for 
conflicting principles has also run into a boundary. I want to see that 
boundary surpassed by my pure approach.

--
robert jasiek
___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] AlphaGo Zero SGF - Free Use or Copyright?

[Xavier Combelle]

what are semantic genetic algorithm ?

to my knowledge genetic algorithm lead to poor result except as a
metaheuristic in optimisation problem


Le 26/10/2017 à 14:40, Jim O'Flaherty a écrit :
> When I get time to spend dozens of hours on computer go again, I plan
> to play in Robert's area with semantic genetic algorithms. I am an
> Architect Software Engineer. Robert's work will allow me better than
> starting entirely from random in much the same way AlphaGo
> bootstrapped from the 100K of professional games. AG0 then leveraged
> AlphaGo in knowing an architecture that was close enough. My intuition
> is my approach will be something similar in it's evolution.
>
> This is the way we're going to "automate" creating provided proofing
> of human cognition styled computer go players to assist humans in a
> gradient ascent learning cycle.
>
> So, Robert, I admire and am encouraged by your research for my own
> computer go projects in this area. Keep kicking butt in your unique
> way. We are in an interesting transition in this community. Stick it
> out. It will be worth it long term.
>
> On Oct 26, 2017 4:38 AM, "Petri Pitkanen"  > wrote:
>
> Unfortunately there is no proof that you principles work better
> than those form eighties. Nor there is any agreement that your
> pronciples form any improvement over the old ones. Yes you are a 
> far better player than me and shows that you are 
> - way better at reading 
> - have hugely better go understanding, principles if you like
>
> What is missing that I doubt that you can verbalise your go
> understanding to degree that by applying those principles  I could
> become substantially better player. again bulleting
> - My reading skills would not get any better hence making much of
> value any learning moot. Obviously issue on me not on your principles
> - your principles are more complex than you understand. Much of
> you know is automated to degree that it is subconsciousness
> information. Transferring that information if hard. Usually done
> by re-playing master games looking at problems i.e. training the
> darn neural net in the head
>
> If you can build Go bot about  KGS 3/4dan strength I am more than
> willing to admit you are right and would even consider buying
> your  books.
>
> Petri
>
> 2017-10-26 6:21 GMT+03:00 Robert Jasiek  >:
>
> On 25.10.2017 18:17, Xavier Combelle wrote:
>
> exact go theory is full of hole.
>
>
> WRT describing the whole game, yes, this is the current state.
> Solving go in a mathematical sense is a project for centuries.
>
> Actually, to my knowledge human can't apply only the exact
> go theory and
> play a decent game.
>
>
> Only for certain positions of a) late endgame, b) semeais, c) ko.
>
> If human can't do that, how it will teach a computer to do
> it magically ?
>
>
> IIRC, Martin Müller implemented CGT endgames a la Mathematical
> Go Endgames.
>
> The reason why (b) had became unpopular is because there
> is no go theory
> precise enough to implement it as an algorithm
>
>
> There is quite some theory of the 95% principle kind which
> might be implemented as approximation. E.g. "Usually, defend
> your weak important group." can be approximated by
> approximating "group", "important" (its loss is too large in a
> quick positional judgement), "weak" (can be killed in two
> successive moves), "defend" (after the move, cannot be killed
> in two successive moves), "usually" (always, unless there are
> several such groups and some must be chosen, say, randomly;
> the approximation being that the alternative strategy of large
> scale exchange is discarded).
>
> Besides, one must prioritise principles to solve conflicting
> principles by a higher order principle.
>
> IMO, such an expert system combined with tree reading and
> maybe MCTS to emulate reading used when a principle depends on
> reading can, with an effort of a few manyears of
> implementation, already achieve amateur mid dan. Not high dan
> yet because high dans can choose advanced strategies, such as
> global exchange, and there are no good enough principles for
> that yet, which would also consider necessary side conditions
> related to influence, aji etc. I need to work out such
> principles during the following years. Currently, the state is
> that weaker principles have identified the major topics
> (influence, aji etc.) to be considered in fights but they must
> be refined to create 95%+ principles.
>
> ***
>
> 

Re: [Computer-go] AlphaGo Zero

[Andy]

I agree with your main point that the first batch of games will be totally
random moves. I just wanted to make a small point that even for totally
random play, the network should be able to learn something about mid-game
positions as well. At move 100, a position with 50 white stones and 40
black stones is likely to be a win for white, even with completely random
play from there, since white has captured 10 black stones.


2017-10-26 8:17 GMT-05:00 Gian-Carlo Pascutto :

> On 25-10-17 16:00, Petr Baudis wrote:
> > That makes sense.  I still hope that with a much more aggressive
> > training schedule we could train a reasonable Go player, perhaps at
> > the expense of worse scaling at very high elos...  (At least I feel
> > optimistic after discovering a stupid bug in my code.)
>
> By the way, a trivial observation: the initial network is random, so
> there's no point in using it for playing the first batch of games. It
> won't do anything useful until it has run a learning pass on a bunch of
> "win/loss" scored games and it can at least tell who is the likely
> winner in the final position (even if it mostly won't be able to make
> territory at first).
>
> This suggests that bootstrapping probably wants 500k starting games with
> just random moves.
>
> FWIW, it does not seem easy to get the value part of the network to
> converge in the dual-res architecture, even when taking the appropriate
> steps (1% weighting on error, strong regularizer).
>
> --
> GCP
> ___
> Computer-go mailing list
> Computer-go@computer-go.org
> http://computer-go.org/mailman/listinfo/computer-go
>
___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] AlphaGo Zero SGF - Free Use or Copyright?

[Robert Jasiek]

On 26.10.2017 08:52, Petri Pitkanen wrote:

Unfortunately there is no proof that you principles work better than those
form eighties.


No computer-go proof.

There is evidence in the form of my playing strength: with the 
principles "from the eighties", I got to circa 1 kyu. L+D reading 
practice etc. made me 3 dan. Afterwards, almost the only thing that made 
me stronger to 5 dan and then further improved my understanding was the 
invention of my own principles.


My principles etc. also work for (an unknown fraction of) readers of my 
books and for a high percentage of my pupils but I cannot compare what 
the effect on them would have been if instead they would only have 
learnt the principles "from the eighties". I do, however, know that my 
principles provide me with very much more efficient means of teaching 
contents compared to using the principles "from the eighties".


The principles "from the eighties" and my principles can be compared 
with each other. IMO, such a comparison is shocking: the principles 
"from the eighties" are very much weaker on average and altogether 
convey very much less contents.



Nor there is any agreement that your pronciples form any
improvement over the old ones.


Only time constraints prevent me from doing an extensive comparison and 
so better support formation of an agreement.



What is missing that I doubt that you can verbalise your go understanding
to degree that by applying those principles  I could become substantially
better player.


Different players are different. So different that some players claim to 
only learn from examples. Therefore, I cannot know whether you are a 
player who could learn well from principles etc.



- My reading skills would not get any better


Do you say so after having learnt and invested effort in applying the 
contents of Tactical Reading?


Regardless of the possible impact of that book, a great part of reading 
skill must be obtained by reading practice in games and problem solving. 
If your reading is much weaker than your knowledge of go theory, then it 
may be the case that almost only reading practise (plus possibly reading 
theory about improving one's reading practice) can significantly improve 
your strength at the moment.



- your principles are more complex than you understand.


I do not think so:)


Much of you know is
automated to degree that it is subconsciousness information.


From ca. 10 kyu to now, especially from 3 dan to now, I have reduced 
the impact of my subconscious thinking on my go decision-making and 
replaced it by knowledge, reading and positional judgement based on 
knowledge and reading. The still remaining subconscious thinking is 
small. Most of my remaining mistakes are related to psychology or 
subconscious thinking, when necessary because of explicit knowledge gaps 
or thinking time constraints.



Transferring that information if hard.


Transferring it from principles etc. to code - yes.


If you can build Go bot about  KGS 3/4dan strength


Using my approach, I expect several manyears, which I do not have for 
that purpose.


--
robert jasiek
___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] AlphaGo Zero SGF - Free Use or Copyright?

[Robert Jasiek]

On 26.10.2017 13:52, Brian Sheppard via Computer-go wrote:

MCTS is the glue that binds incompatible rules.


This is, however, not what I mean. Conflicting principles (call them 
rules if you like) must be dissolved by higher order principles. Only 
when all conflicts are dissolved should MCTS be applied.


What you describe has been used with success and better success than I 
expect what my knowledge-pure approach can currently achieve. But MCTS 
as glue for conflicting principles has also run into a boundary. I want 
to see that boundary surpassed by my pure approach.


--
robert jasiek
___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] AlphaGo Zero

[Gian-Carlo Pascutto]

On 25-10-17 16:00, Petr Baudis wrote:
> That makes sense.  I still hope that with a much more aggressive 
> training schedule we could train a reasonable Go player, perhaps at
> the expense of worse scaling at very high elos...  (At least I feel 
> optimistic after discovering a stupid bug in my code.)

By the way, a trivial observation: the initial network is random, so
there's no point in using it for playing the first batch of games. It
won't do anything useful until it has run a learning pass on a bunch of
"win/loss" scored games and it can at least tell who is the likely
winner in the final position (even if it mostly won't be able to make
territory at first).

This suggests that bootstrapping probably wants 500k starting games with
just random moves.

FWIW, it does not seem easy to get the value part of the network to
converge in the dual-res architecture, even when taking the appropriate
steps (1% weighting on error, strong regularizer).

-- 
GCP
___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] AlphaGo Zero SGF - Free Use or Copyright?

[Jim O'Flaherty]

When I get time to spend dozens of hours on computer go again, I plan to
play in Robert's area with semantic genetic algorithms. I am an Architect
Software Engineer. Robert's work will allow me better than starting
entirely from random in much the same way AlphaGo bootstrapped from the
100K of professional games. AG0 then leveraged AlphaGo in knowing an
architecture that was close enough. My intuition is my approach will be
something similar in it's evolution.

This is the way we're going to "automate" creating provided proofing of
human cognition styled computer go players to assist humans in a gradient
ascent learning cycle.

So, Robert, I admire and am encouraged by your research for my own computer
go projects in this area. Keep kicking butt in your unique way. We are in
an interesting transition in this community. Stick it out. It will be worth
it long term.

On Oct 26, 2017 4:38 AM, "Petri Pitkanen" 
wrote:

> Unfortunately there is no proof that you principles work better than those
> form eighties. Nor there is any agreement that your pronciples form any
> improvement over the old ones. Yes you are a  far better player than me and
> shows that you are
> - way better at reading
> - have hugely better go understanding, principles if you like
>
> What is missing that I doubt that you can verbalise your go understanding
> to degree that by applying those principles  I could become substantially
> better player. again bulleting
> - My reading skills would not get any better hence making much of value
> any learning moot. Obviously issue on me not on your principles
> - your principles are more complex than you understand. Much of you know
> is automated to degree that it is subconsciousness information.
> Transferring that information if hard. Usually done by re-playing master
> games looking at problems i.e. training the darn neural net in the head
>
> If you can build Go bot about  KGS 3/4dan strength I am more than willing
> to admit you are right and would even consider buying your  books.
>
> Petri
>
> 2017-10-26 6:21 GMT+03:00 Robert Jasiek :
>
>> On 25.10.2017 18:17, Xavier Combelle wrote:
>>
>>> exact go theory is full of hole.
>>>
>>
>> WRT describing the whole game, yes, this is the current state. Solving go
>> in a mathematical sense is a project for centuries.
>>
>> Actually, to my knowledge human can't apply only the exact go theory and
>>> play a decent game.
>>>
>>
>> Only for certain positions of a) late endgame, b) semeais, c) ko.
>>
>> If human can't do that, how it will teach a computer to do it magically ?
>>>
>>
>> IIRC, Martin Müller implemented CGT endgames a la Mathematical Go
>> Endgames.
>>
>> The reason why (b) had became unpopular is because there is no go theory
>>> precise enough to implement it as an algorithm
>>>
>>
>> There is quite some theory of the 95% principle kind which might be
>> implemented as approximation. E.g. "Usually, defend your weak important
>> group." can be approximated by approximating "group", "important" (its loss
>> is too large in a quick positional judgement), "weak" (can be killed in two
>> successive moves), "defend" (after the move, cannot be killed in two
>> successive moves), "usually" (always, unless there are several such groups
>> and some must be chosen, say, randomly; the approximation being that the
>> alternative strategy of large scale exchange is discarded).
>>
>> Besides, one must prioritise principles to solve conflicting principles
>> by a higher order principle.
>>
>> IMO, such an expert system combined with tree reading and maybe MCTS to
>> emulate reading used when a principle depends on reading can, with an
>> effort of a few manyears of implementation, already achieve amateur mid
>> dan. Not high dan yet because high dans can choose advanced strategies,
>> such as global exchange, and there are no good enough principles for that
>> yet, which would also consider necessary side conditions related to
>> influence, aji etc. I need to work out such principles during the following
>> years. Currently, the state is that weaker principles have identified the
>> major topics (influence, aji etc.) to be considered in fights but they must
>> be refined to create 95%+ principles.
>>
>> ***
>>
>> In the 80s and 90s, expert systems failed to do better than ca. 5 kyu
>> because principles were only marginally better than 50%. Today, (my)
>> average principles discard the weaker, 50% principles and are ca. 75%.
>> Tomorrow, the 75% principles can be discarded for an average of 95%
>> principles. Expert systems get their chance again! Their major disadvantage
>> remains: great manpower is required for implementation. The advantage is
>> semantical understanding.
>>
>> --
>> robert jasiek
>>
>> ___
>> Computer-go mailing list
>> Computer-go@computer-go.org
>> http://computer-go.org/mailman/listinfo/computer-go
>>
>
>
> 

Re: [Computer-go] AlphaGo Zero SGF - Free Use or Copyright?

[Brian Sheppard via Computer-go]

Robert is right, but Robert seems to think this hasn't been done. Actually 
every prominent non-neural MCTS program since Mogo has been based on the exact 
design that Robert describes. The best of them achieve somewhat greater 
strength than Robert expects.

MCTS is the glue that binds incompatible rules. It rationalizes different 
heuristics into a coherent whole by testing the ideas in a competition against 
one another using a meaningful evaluation (win/loss).

Best,
Brian

-Original Message-
From: Computer-go [mailto:computer-go-boun...@computer-go.org] On Behalf Of 
Xavier Combelle
Sent: Thursday, October 26, 2017 1:50 AM
To: computer-go@computer-go.org
Subject: Re: [Computer-go] AlphaGo Zero SGF - Free Use or Copyright?



>> The reason why (b) had became unpopular is because there is no go 
>> theory precise enough to implement it as an algorithm
>
> There is quite some theory of the 95% principle kind which might be 
> implemented as approximation. E.g. "Usually, defend your weak 
> important group." can be approximated by approximating "group", 
> "important" (its loss is too large in a quick positional judgement), 
> "weak" (can be killed in two successive moves), "defend" (after the 
> move, cannot be killed in two successive moves), "usually" (always, 
> unless there are several such groups and some must be chosen, say, 
> randomly; the approximation being that the alternative strategy of 
> large scale exchange is discarded).
>
> Besides, one must prioritise principles to solve conflicting 
> principles by a higher order principle.
>
> IMO, such an expert system combined with tree reading and maybe MCTS 
> to emulate reading used when a principle depends on reading can, with 
> an effort of a few manyears of implementation, already achieve amateur 
> mid dan. Not high dan yet because high dans can choose advanced 
> strategies, such as global exchange, and there are no good enough 
> principles for that yet, which would also consider necessary side 
> conditions related to influence, aji etc. I need to work out such 
> principles during the following years. Currently, the state is that 
> weaker principles have identified the major topics (influence, aji
> etc.) to be considered in fights but they must be refined to create 
> 95%+ principles.
>
> ***
>
> In the 80s and 90s, expert systems failed to do better than ca. 5 kyu 
> because principles were only marginally better than 50%. Today, (my) 
> average principles discard the weaker, 50% principles and are ca. 75%.
> Tomorrow, the 75% principles can be discarded for an average of 95% 
> principles. Expert systems get their chance again! Their major 
> disadvantage remains: great manpower is required for implementation.
> The advantage is semantical understanding.
>
From a software developer point of view enlighten by my knowledge of history of 
ai and history of go development,
 such approximate definition is close to useless to build a software at the 
current state of art.
One of the reason is as you state the considerable work it would require to 
implement a huge number of imprecise rules.
As you are not a software developer, I want you to look on this comics which 
state the difference between apparent difficulty and real difficulty of 
developping software. https://xkcd.com/1425/ As far as I understand your task 
to implement such an expert system would require the many years of 
implementations would be thousands of years.
As far as my experience speak the expected reward would be a win of one or two 
rank and so definitely not a mid dan amateur level.

Xavier Combelle

___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] AlphaGo Zero SGF - Free Use or Copyright?

[Petri Pitkanen]

Unfortunately there is no proof that you principles work better than those
form eighties. Nor there is any agreement that your pronciples form any
improvement over the old ones. Yes you are a  far better player than me and
shows that you are
- way better at reading
- have hugely better go understanding, principles if you like

What is missing that I doubt that you can verbalise your go understanding
to degree that by applying those principles  I could become substantially
better player. again bulleting
- My reading skills would not get any better hence making much of value any
learning moot. Obviously issue on me not on your principles
- your principles are more complex than you understand. Much of you know is
automated to degree that it is subconsciousness information. Transferring
that information if hard. Usually done by re-playing master games looking
at problems i.e. training the darn neural net in the head

If you can build Go bot about  KGS 3/4dan strength I am more than willing
to admit you are right and would even consider buying your  books.

Petri

2017-10-26 6:21 GMT+03:00 Robert Jasiek :

> On 25.10.2017 18:17, Xavier Combelle wrote:
>
>> exact go theory is full of hole.
>>
>
> WRT describing the whole game, yes, this is the current state. Solving go
> in a mathematical sense is a project for centuries.
>
> Actually, to my knowledge human can't apply only the exact go theory and
>> play a decent game.
>>
>
> Only for certain positions of a) late endgame, b) semeais, c) ko.
>
> If human can't do that, how it will teach a computer to do it magically ?
>>
>
> IIRC, Martin Müller implemented CGT endgames a la Mathematical Go Endgames.
>
> The reason why (b) had became unpopular is because there is no go theory
>> precise enough to implement it as an algorithm
>>
>
> There is quite some theory of the 95% principle kind which might be
> implemented as approximation. E.g. "Usually, defend your weak important
> group." can be approximated by approximating "group", "important" (its loss
> is too large in a quick positional judgement), "weak" (can be killed in two
> successive moves), "defend" (after the move, cannot be killed in two
> successive moves), "usually" (always, unless there are several such groups
> and some must be chosen, say, randomly; the approximation being that the
> alternative strategy of large scale exchange is discarded).
>
> Besides, one must prioritise principles to solve conflicting principles by
> a higher order principle.
>
> IMO, such an expert system combined with tree reading and maybe MCTS to
> emulate reading used when a principle depends on reading can, with an
> effort of a few manyears of implementation, already achieve amateur mid
> dan. Not high dan yet because high dans can choose advanced strategies,
> such as global exchange, and there are no good enough principles for that
> yet, which would also consider necessary side conditions related to
> influence, aji etc. I need to work out such principles during the following
> years. Currently, the state is that weaker principles have identified the
> major topics (influence, aji etc.) to be considered in fights but they must
> be refined to create 95%+ principles.
>
> ***
>
> In the 80s and 90s, expert systems failed to do better than ca. 5 kyu
> because principles were only marginally better than 50%. Today, (my)
> average principles discard the weaker, 50% principles and are ca. 75%.
> Tomorrow, the 75% principles can be discarded for an average of 95%
> principles. Expert systems get their chance again! Their major disadvantage
> remains: great manpower is required for implementation. The advantage is
> semantical understanding.
>
> --
> robert jasiek
>
> ___
> Computer-go mailing list
> Computer-go@computer-go.org
> http://computer-go.org/mailman/listinfo/computer-go
>
___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] AlphaGo Zero SGF - Free Use or Copyright?

[Xavier Combelle]

>> The reason why (b) had became unpopular is because there is no go theory
>> precise enough to implement it as an algorithm
>
> There is quite some theory of the 95% principle kind which might be
> implemented as approximation. E.g. "Usually, defend your weak
> important group." can be approximated by approximating "group",
> "important" (its loss is too large in a quick positional judgement),
> "weak" (can be killed in two successive moves), "defend" (after the
> move, cannot be killed in two successive moves), "usually" (always,
> unless there are several such groups and some must be chosen, say,
> randomly; the approximation being that the alternative strategy of
> large scale exchange is discarded).
>
> Besides, one must prioritise principles to solve conflicting
> principles by a higher order principle.
>
> IMO, such an expert system combined with tree reading and maybe MCTS
> to emulate reading used when a principle depends on reading can, with
> an effort of a few manyears of implementation, already achieve amateur
> mid dan. Not high dan yet because high dans can choose advanced
> strategies, such as global exchange, and there are no good enough
> principles for that yet, which would also consider necessary side
> conditions related to influence, aji etc. I need to work out such
> principles during the following years. Currently, the state is that
> weaker principles have identified the major topics (influence, aji
> etc.) to be considered in fights but they must be refined to create
> 95%+ principles.
>
> ***
>
> In the 80s and 90s, expert systems failed to do better than ca. 5 kyu
> because principles were only marginally better than 50%. Today, (my)
> average principles discard the weaker, 50% principles and are ca. 75%.
> Tomorrow, the 75% principles can be discarded for an average of 95%
> principles. Expert systems get their chance again! Their major
> disadvantage remains: great manpower is required for implementation.
> The advantage is semantical understanding.
>
From a software developer point of view enlighten by my knowledge of
history of ai and history of go development,
 such approximate definition is close to useless to build a software at
the current state of art.
One of the reason is as you state the considerable work it would require
to implement a huge number of imprecise rules.
As you are not a software developer, I want you to look on this comics
which state the difference between apparent difficulty and real difficulty
of developping software. https://xkcd.com/1425/
As far as I understand your task to implement such an expert system
would require the many years of implementations would be thousands of years.
As far as my experience speak the expected reward would be a win of one
or two rank and so definitely not a mid dan amateur level.

Xavier Combelle

___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] AlphaGo Zero SGF - Free Use or Copyright?

[Darren Cook]

> What do you want evaluate the software for ? corner cases which never
> have happen in a real game ?

If the purpose of this mailing list is a community to work out how to
make a 19x19 go program that can beat any human, then AlphaGo has
finished the job, and we can shut it down.

But this list has always been for anything related to computers and the
game of go. Right from John Tromp counting the number of games through
to tips and hints on the best compiler flags to use.

BTW, I noticed in the paper that it showed 3 games AlphaGo Zero lost to
AlphaGo Master: in game 11 Zero had white, in games 14 and 16 Zero had
black. An opponent that can only win 11% of games against it, was able
to win on both sides of the komi. Suggesting there is still quite a bit
of room for improvement.

Darren
___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] AlphaGo Zero

[Gian-Carlo Pascutto]

On 25-10-17 16:00, Petr Baudis wrote:

>> The original paper has the value they used. But this likely needs tuning. I
>> would tune with a supervised network to get started, but you need games for
>> that. Does it even matter much early on? The network is random :)
> 
>   The network actually adapts quite rapidly initially, in my experience.
> (Doesn't mean it improves - it adapts within local optima of the few
> games it played so far.)

Yes, but once there's structure, you can tune the parameter with CLOP or
whatever.

>   Yes, but why wouldn't you want that randomness in the second or third
> move?

You only need to play a different move at the root in order for the game
to deviate.

-- 
GCP
___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] AlphaGo Zero SGF - Free Use or Copyright?

[Xavier Combelle]

Le 24/10/2017 à 22:41, Robert Jasiek a écrit :

> On 24.10.2017 20:19, Xavier Combelle wrote:
>> totally unrelated
>
> No, because a) software must also be evaluated and can by go theory and
What do you want evaluate the software for ? corner cases which never
have happen in a real game ?

The current testing way that deepmind used
that is: first of all making software and software-human tournament,
guess pro move, guess pro game result
was amply enough to make the best go software.

> b) software can be built on exact go theory. That currently (b) is
> unpopular does not mean unrelated.
>
It is just a wild guess. exact go theory is full of hole.
Actually, to my knowledge human can't apply only the exact go theory and
play a decent game.
If human can't do that, how it will teach a computer to do it magically ?

if you want we can setup a game were you apply only exact go theory
against me (I'm only 2 kyu)
The rules are the following, you have to apply mechanically the go
theory as a computer would do
at each move such as I could do exactly the same
and show in a detailed way how you applied it. If you won I will
recognize  the fact that the exact go
theory is not full of hole.

The reason why (b) had became unpopular is because there is no go theory
precise enough to implement it as an algorithm
and MCTS and neural network was way to use small or none part of go
theory and make a decent player.

___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] AlphaGo Zero

[Petr Baudis]

On Fri, Oct 20, 2017 at 08:02:02PM +, Gian-Carlo Pascutto wrote:
> On Fri, Oct 20, 2017, 21:48 Petr Baudis  wrote:
> 
> >   Few open questions I currently have, comments welcome:
> >
> >   - there is no input representing the number of captures; is this
> > information somehow implicit or can the learned winrate predictor
> > never truly approximate the true values because of this?
> >
> 
> They are using Chinese rules, so prisoners don't matter. There are simply
> less stones of one color on the board.

  Right!  No idea what was I thinking.

> >   - what ballpark values for c_{puct} are reasonable?
> >
> 
> The original paper has the value they used. But this likely needs tuning. I
> would tune with a supervised network to get started, but you need games for
> that. Does it even matter much early on? The network is random :)

  The network actually adapts quite rapidly initially, in my experience.
(Doesn't mean it improves - it adapts within local optima of the few
games it played so far.)

> >   - why is the dirichlet noise applied only at the root node, if it's
> > useful?
> >
> 
> It's only used to get some randomness in the move selection, no ? It's not
> actually useful for anything besides that.

  Yes, but why wouldn't you want that randomness in the second or third
move?

> >   - the training process is quite lazy - it's not like the network sees
> > each game immediately and adjusts, it looks at last 500k games and
> > samples 1000*2048 positions, meaning about 4 positions per game (if
> > I understood this right) - I wonder what would happen if we trained
> > it more aggressively, and what AlphaGo does during the initial 500k
> > games; currently, I'm training on all positions immediately, I guess
> > I should at least shuffle them ;)
> >
> 
> I think the lazyness may be related to the concern that reinforcement
> methods can easily "forget" things they had learned before. The value
> network training also likes positions from distinct games.

  That makes sense.  I still hope that with a much more aggressive
training schedule we could train a reasonable Go player, perhaps at the
expense of worse scaling at very high elos...  (At least I feel
optimistic after discovering a stupid bug in my code.)

-- 
Petr Baudis, Rossum
Run before you walk! Fly before you crawl! Keep moving forward!
If we fail, I'd rather fail really hugely.  -- Moist von Lipwig
___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] AlphaGo Zero SGF - Free Use or Copyright?

[uurtamo .]

We're suffering under the burden of so much success from other methods that
​it's hard for many people to imagine that anything else is worth
considering.

Of course this is not true.

Tromp's enumerations are particularly enjoyable for me.

Human-built decision trees have been so unsuccessful, compared with
machine-learned models, for around 25 years, that only a few tiny wisps of
academia are interested in them in a serious way that industry can and
should take seriously.

Some control-system methods, some ILP, some NLP, etc., are all successful
counterexamples, in many cases in the field of logistics, transportation,
etc. Complicated games such as go have pretty much not fallen due to these
methods.

(As a coworker of mine said recently, "It's probably going to be okay to
hard-code the rule for the self-driving car not to hit pedestrians; there's
no need to train with lots of examples of hitting pedestrians to train your
algorithm".)

They (analytically exact methods) are still interesting to study from a
game-theoretic persepective, mathematically. There are exact solvers for
all kinds of specialized problems.

Problems with more than a few variables can very easily lead to many or
most cases not being exactly (analytically) soluble. That's why all of
these probabilistic approximation methods are so successful. They don't
have to be exactly right. It's easing the constraint most people care least
about (exact certitude of a win or success locally rather than an extremely
high probability of a win or success locally).

Asking the "high probability of success" guys to explain why their method
works is a particularly galling (and trite) way of messing with them. They
can just point to the results. The reason is that they don't know. And it's
going to be a very, very long time before they do.

At a fundamental level, probabilistic methods seem to be (some
theoreticians believe) more powerful than non-probabilistic methods for
relatively hard (as opposed to very very hard) problems. This is nicely
encoded by computational complexity theorists as the question BPP = P ?

steve



On Tue, Oct 24, 2017 at 1:41 PM, Robert Jasiek  wrote:

> On 24.10.2017 20:19, Xavier Combelle wrote:
>
>> totally unrelated
>>
>
> No, because a) software must also be evaluated and can by go theory and b)
> software can be built on exact go theory. That currently (b) is unpopular
> does not mean unrelated.
>
> --
> robert jasiek
>
> ___
> Computer-go mailing list
> Computer-go@computer-go.org
> http://computer-go.org/mailman/listinfo/computer-go
>
___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] AlphaGo Zero SGF - Free Use or Copyright?

[Robert Jasiek]

On 24.10.2017 20:19, Xavier Combelle wrote:

totally unrelated


No, because a) software must also be evaluated and can by go theory and 
b) software can be built on exact go theory. That currently (b) is 
unpopular does not mean unrelated.


--
robert jasiek
___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] AlphaGo Zero SGF - Free Use or Copyright?

[Darren Cook]

Could we PLEASE take this off-list? If you don't like someone, or what
they post, filter them. If you think someone should be banned, present
your case to the list owner(s).

Darren

___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] AlphaGo Zero SGF - Free Use or Copyright?

[Xavier Combelle]

"In the current time, computer-go discussion and research has a very
high percentage of people discussing the side of mainly programs and
programming but I belong to the very low percentage of people
discussing mainly go-theoretical aspects of computer-go. With a higher
percentage of the latter, there would also be more discussions
resulting to something."

Now you explained what you describe what you mean by go-theoretical
aspects, which is your main area of interest,
I feel like they are totally unrelated to the purpose of this mailing list.

to quote the home page: http://computer-go.org/
"computer-go: Discussion on research and development of software that
plays the game of Go."

Now that is clear, I understand why I always felt your intervention
misplaced (because they were misplaced).

Le 24/10/2017 à 17:00, Robert Jasiek a écrit :
> On 24.10.2017 16:45, Xavier Combelle wrote:
>> I don't understand what you mean by go-theorical aspects.
>
> Go theory is an ambiguous term and means everything from informal
> ("Starting with a standard corner move can't be wrong.") via principle
> ("Usually, defend a weak important group.") to formal (
> https://senseis.xmp.net/?CycleLaw ).
>
>> and especially when applying to computer-go.
>
> Relating computer play / algorithms to go theory or vice versa adds
> another layer of difficulty indeed.
>
>> To my knowledge the only theoretical (in a
>> mathematic meaning of theoretical) approach of go is combinatorial
>> theory and it leads to very few knowledge.
>
> Other mathematical theory with practical relevance is related to
> capturing races (see Capturing Races 1 - Two Basic Groups, Thomas
> Wolf's papers etc., endgame (e.g.,
> http://home.snafu.de/jasiek/kodame.pdf and google for related proofs)
> or will be published by me later (will be quite a lot and have
> practical relevance, but you need to be patient). Research in
> mathematical go theory requires much time because exactness is often
> necessary and proving can be tricky.
>


___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] AlphaGo Zero SGF - Free Use or Copyright?

[Robert Jasiek]

On 24.10.2017 16:45, Xavier Combelle wrote:

I don't understand what you mean by go-theorical aspects.


Go theory is an ambiguous term and means everything from informal 
("Starting with a standard corner move can't be wrong.") via principle 
("Usually, defend a weak important group.") to formal ( 
https://senseis.xmp.net/?CycleLaw ).



and especially when applying to computer-go.


Relating computer play / algorithms to go theory or vice versa adds 
another layer of difficulty indeed.



To my knowledge the only theoretical (in a
mathematic meaning of theoretical) approach of go is combinatorial
theory and it leads to very few knowledge.


Other mathematical theory with practical relevance is related to 
capturing races (see Capturing Races 1 - Two Basic Groups, Thomas Wolf's 
papers etc., endgame (e.g., http://home.snafu.de/jasiek/kodame.pdf and 
google for related proofs) or will be published by me later (will be 
quite a lot and have practical relevance, but you need to be patient). 
Research in mathematical go theory requires much time because exactness 
is often necessary and proving can be tricky.


--
robert jasiek
___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] AlphaGo Zero SGF - Free Use or Copyright?

[Xavier Combelle]

Le 24/10/2017 à 14:35, Robert Jasiek a écrit :
> On 24.10.2017 11:45, David Ongaro wrote:
>> very seldom saw a discussion with Robert lead to anything.
>
> (You seem to only refer to discussion on this mailing list.)
>
> Apart from this being a discussion about one particular person, let me
> ignore this for a moment:
>
> In the current time, computer-go discussion and research has a very
> high percentage of people discussing the side of mainly programs and
> programming but I belong to the very low percentage of people
> discussing mainly go-theoretical aspects of computer-go. With a higher
> percentage of the latter, there would also be more discussions
> resulting to something.
>
I don't understand what you mean by go-theorical aspects. and especially
when applying to computer-go. To my knowledge the only theoretical (in a
mathematic meaning of theoretical) approach of go is combinatorial
theory and it leads to very few knowledge. Can you explain what you mean
maybe by giving example ?
___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] AlphaGo Zero SGF - Free Use or Copyright?

[Jim O'Flaherty]

David Ongaro and Xavier Combelle,

I am respectfully requesting you stop inappropriately discussing and
addressing the person Robert Jasiek in your posts. He has not acted in any
way inappropriate on this list (I fully read every post). Therefore he
hasn't done anything which needs to be addressed regarding his
participation. However, each of you are acting inappropriately. Neither of
you are the final arbiter of what is valuable and/or appropriate for dialog
on this forum. And each of you has wandered into the space of inappropriate
discussion of a contributor here.

I enjoy Robert's posts. All of them. Yes, that includes the ones about
which each of you are complaining. Just because you do not value them
doesn't mean I don't value them. And, I also know there are others who ALSO
value Robert's posts. All of them.

As was said in an earlier reply, your email has a simple filtering
function. If you do not like a particular person's posts to this email
list, simply add their email to your list of blocked/ignored so it goes to
your spam or trash buckets and you never see it. IOW, please take
responsibility for your character and behavior and refrain from posting
non-Go related diatribes ESPECIALLY about other participating members.


Respectfully,

Jim O'Flaherty


On Tue, Oct 24, 2017 at 5:42 AM,  wrote:

> On 2017-10-23 at 23:56, Thomas Rohde  wrote:
>
> > On 2017-10-23 at 19:15, Xavier Combelle 
> wrote:
> >
> > > Hi Robert Jasiek,
> > >
> > > you might have a delusional way to see the game of go and life,
> >
> > this is quite an insult
>
> Do you consider Robert's style of discussion "kind"? I for my part do not.
>
> I'm not saying that Robert's research in the area of Go corner cases
> doesn't have any value, it certainly has. One probably needs a certain kind
> of dedication to do it. But trying to bend every topic into this area is
> more often than not uncalled for.
>
> I don't know what it is. Maybe it's a certain kind of arrogance, resulting
> from the fact of knowing more than anybody else about a certain area in Go.
> But in the end it doesn't matter what it is, we all have our faults. What
> matters is that I very seldom saw a discussion with Robert lead to anything.
>
> David
>
>
> ___
> Computer-go mailing list
> Computer-go@computer-go.org
> http://computer-go.org/mailman/listinfo/computer-go
>
___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] AlphaGo Zero SGF - Free Use or Copyright?

[Robert Jasiek]

On 24.10.2017 11:45, David Ongaro wrote:

very seldom saw a discussion with Robert lead to anything.


(You seem to only refer to discussion on this mailing list.)

Apart from this being a discussion about one particular person, let me 
ignore this for a moment:


In the current time, computer-go discussion and research has a very high 
percentage of people discussing the side of mainly programs and 
programming but I belong to the very low percentage of people discussing 
mainly go-theoretical aspects of computer-go. With a higher percentage 
of the latter, there would also be more discussions resulting to something.


--
robert jasiek
___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] AlphaGo Zero SGF - Free Use or Copyright?

[david . ongaro]

On 2017-10-23 at 23:56, Thomas Rohde  wrote:
> On 2017-10-23 at 19:15, Xavier Combelle  wrote:
>
> > Hi Robert Jasiek,
> > 
> > you might have a delusional way to see the game of go and life,
> 
> this is quite an insult
Do you consider Robert's style of discussion "kind"? I for my part do
not.

I'm not saying that Robert's research in the area of Go corner cases
doesn't have any value, it certainly has. One probably needs a certain
kind of dedication to do it. But trying to bend every topic into this
area is more often than not uncalled for.

I don't know what it is. Maybe it's a certain kind of arrogance,
resulting from the fact of knowing more than anybody else about a
certain area in Go. But in the end it doesn't matter what it is, we
all have our faults. What matters is that I very seldom saw a
discussion with Robert lead to anything.

David

___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] AlphaGo Zero SGF - Free Use or Copyright?

[Robert Jasiek]

On 23.10.2017 19:15, Xavier Combelle wrote:

[personal attack deleted]
Did you already encounter a real game with "disturbing life kos or
anti-sekis" and especially "ladders (...) beyond 250 moves" ? If not how
do you believe that Alphago would learn how to manage such situations.


Dave Dyer wrote:

I wonder how alphago-0 treats the menagerie of special positions, such as
bent 4 in the corner, thousand year ko, rotating ko, etc.


uurtamo wrote:
> It will be interesting to realize that those specialized positions
> (thousand-year-ko, bent 4) are actually a microscopic issue in
> game-winning.

The exceptional cases may be rarities in practical play but not all are 
that rare. E.g., I have had two games in roughly 40,000 ending with a 
double ko seki. Already one "rarity" occurring occasionally means that 
all rarities occur more often in practice. Therefore they do have 
practical relevance. Quite like a white truck is relevant and not to be 
confused with the sky, or an AI car can kill (which has happened because 
of such a "rarity"). In go, the consequences or misjudging "rarities" 
are just up a lost game, but this is the very purpose - avoiding lost 
games by avoiding errors. Rarities are good test samples for checking 
whether an AI program avoids errors in non-standard situations.


The same must be studied for standard situations, whose deeper details 
can also lead to errors. Not because a standard by itself would be 
difficult but because the deeper details increase complexity and this 
can lead to errors. Studying the standards and identifying errors in 
their deeper details can be difficult. E.g., we see AlphaGo (Zero) 
invading and living in a large moyo or not invading and wonder why. Part 
of the answer would be: invading and living is impossible. Studying this 
is complex because it involves deep reading for the standard case of a 
moyo and the question of invading it.


The "rarities" are infrequent but can be good test tools because 
distinguishing correct from wrong play can be easy if a rarity's 
behaviour is understood well. The standards are frequent but often not 
the best test tools because many standards interact with each other and 
they all depend on deep reading and exact positional judgement.


I cannot know if AlphaGo Zero has already learnt how to play in (some) 
rarities (those that can be solved earlier than the constant game end 
rule; e.g., we cannot test 4 octuple kos), will learn it or would not be 
able to learn it - but I want to know. In particular, because I want to 
know which errors AlphaGo Zero does make.


I want to know this for go and for the general AI project. Avoiding 
errors is essential for both. I do not fall into the illusion that 
AlphaGo Zero would be the perfect player but expect that it can make 
errors at any (unexpected) time. We need to understand what causes 
errors, how frequent they are and what most extreme consequences they 
can have. Rarities are one very good study tool for this purpose.


--
robert jasiek
___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] Alphago Zero special circumstances

[Ray Tayek]

On 10/23/2017 11:29 AM, Dave Dyer wrote:


I wonder how alphago-0 treats the menagerie of special positions, such as
bent 4 in the corner, thousand year ko, rotating ko, etc.



they are using chinese rules, so bent four just gets played out. have no 
idea about strange ko's. there was a double ko in one of the games.


i looked at the 20 games that 0 played against master. there are a bunch 
of cases where the star point gets invaded very early in the game. 
sometimes by both players. and the direction that the opponent uses to 
block is counter intuitive. the older alphago made a lot of "kosumi 
approach" moves. this the same shape when invading the star point, so i 
thought that this was strange.


mr. yang's answer to this was that the corner is important/valuable.

maybe there a lot more to learn about the game.

thanks


--
Honesty is a very expensive gift. So, don't expect it from cheap people 
- Warren Buffett

http://tayek.com/
___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] Alphago Zero special circumstances

[Hideki Kato]

Zero's MCTS code knows "legal" moves (as a part of rules) 
embed (page 22).  Bent four will be solved in practice as 
she uses Chinese rules.

Hideki

Dave Dyer: <20171023210343.e65bc31a...@eugeneweb.com>:
>
>I wonder how alphago-0 treats the menagerie of special positions, such as
>bent 4 in the corner, thousand year ko, rotating ko, etc.  
>
>___
>Computer-go mailing list
>Computer-go@computer-go.org
>http://computer-go.org/mailman/listinfo/computer-go
-- 
Hideki Kato 
___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] AlphaGo Zero SGF - Free Use or Copyright?

[Thomas Rohde]

On 2017-10-23 at 19:15, Xavier Combelle  wrote:

> Hi Robert Jasiek,
> 
> you might have a delusional way to see the game of go and life,

this is quite an insult, IMO, and I’d prefer not to read such personal attacks 
in this list. What about staying on-topic? 


> but I would love that you would not pollute my mailbox with such a delusional 
> vision.

May I suggest that you simply create a frigin’ filter? Also, you always have 
the liberty to hit the delete or backspace key, I’m sure you know where on your 
keyboard they are ;-)


> I'm certain that a lot of person of this mailing list and other forums share 
> my view.

FTR: Not me.


> To sum up, I would be pleased and I'm quite certain others too that you 
> consider seriously behave more like others persons.

LOL, “like other persons” :-D
What a boring list this would be if everybody would be “like other persons”.


Respectfully,

Tom


Great minds discuss ideas; average minds discuss events; small minds discuss 
people.
— Eleanor Roosevelt (though the authorship is disputed)


-- 
Thomas Rohde
Wiesenkamp 12, 29646 Bispingen, GERMANY
--
t...@bonobo.com
___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] Alphago Zero special circumstances

[uurtamo .]

It will be interesting to realize that those specialized positions
(thousand-year-ko, bent 4) are actually a microscopic issue in game-winning.

s.


On Mon, Oct 23, 2017 at 2:10 PM, uurtamo .  wrote:

> We can all "wonder" such things unless we are not too busy to build some
> code to filter out such positions and see what actually happened in the
> self-play games opened up to everyone to see.
>
> s.
>
>
> On Mon, Oct 23, 2017 at 11:29 AM, Dave Dyer  wrote:
>
>>
>> I wonder how alphago-0 treats the menagerie of special positions, such as
>> bent 4 in the corner, thousand year ko, rotating ko, etc.
>>
>> ___
>> Computer-go mailing list
>> Computer-go@computer-go.org
>> http://computer-go.org/mailman/listinfo/computer-go
>
>
>
___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] Alphago Zero special circumstances

[uurtamo .]

We can all "wonder" such things unless we are not too busy to build some
code to filter out such positions and see what actually happened in the
self-play games opened up to everyone to see.

s.


On Mon, Oct 23, 2017 at 11:29 AM, Dave Dyer  wrote:

>
> I wonder how alphago-0 treats the menagerie of special positions, such as
> bent 4 in the corner, thousand year ko, rotating ko, etc.
>
> ___
> Computer-go mailing list
> Computer-go@computer-go.org
> http://computer-go.org/mailman/listinfo/computer-go
___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] Alphago Zero special circumstances

[Dave Dyer]

I wonder how alphago-0 treats the menagerie of special positions, such as
bent 4 in the corner, thousand year ko, rotating ko, etc.  

___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] Alphago Zero special circumstances

[Dave Dyer]

I wonder how alphago-0 treats the menagerie of special positions, such as
bent 4 in the corner, thousand year ko, rotating ko, etc.  

___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] AlphaGo Zero SGF - Free Use or Copyright?

[Jim O'Flaherty]

If you wanted to do research on a specific joseki to see if AG0 found
weaknesses or explored alternatives, especially in games past day 3 (when
it began besting AGM), it would be quite interesting to see, through that
explicit filter, what kinds of things emerged around that specific joseki.
You might even discover that the joseki has weaknesses in several early
branches which haven't been found by humans in hundreds of years of using
the joseki.

Another angle is pretty much every game from AG0 day 4 forward would be
superior to the 100,000 games of the pro players upon which AGM was
trained. IOW, that is a huge number of games that exceed the very best
human-vs-human games quality and would be quite superior to seeding with
those human (biased) games for any other engine playing in the area.

Or, someone training a net, could walk up the 29 million games and explore
how their engine differs, make adjustments and explore climbing higher.
This data set enables "higher" in a totally different way than any prior
data set.

The number and variety of things that could be explored with the 29 million
games boggles the mind. I am deeply hopeful they do some form of simple
compressed dump of it somewhere. Then, it would be just a matter of getting
it loaded into a DB to create all sorts of indexes, queries, novel subsets,
etc.


Namaste,

Jim O'Flaherty
Founder/CEO
Precision Location Intelligence, Inc.
 • Irving, TX, USA
469-358-0633 <4693580633> • jim.oflaherty...@gmail.com •
www.linkedin.com/in/jimoflahertyjr

CONFIDENTIALITY / PROPRIETARY NOTICE:
The information contained in this e-mail, including any attachment(s), is
confidential information for Precision Location Intelligence, Inc.. As
such, it may be privileged and exempt from disclosure under applicable law.
If the reader of this message is not the intended recipient, or if you
received this message in error, then any direct or indirect disclosure,
distribution or copying of this message is strictly prohibited. If you have
received this message in error, please notify Precision Location
Intelligence, Inc. by calling (214) 489-7779 <2144897779> immediately and
by sending a return e-mail; delete this message; and destroy all copies,
including attachments.

On Mon, Oct 23, 2017 at 9:44 AM, Petri Pitkanen 
wrote:

> If the AG got better by playing against itself rather than training on
> previous good players then I do not thing training data is that important.
> Perhaps it is but google has shown that actually u dont need it. Just loads
> of processing will do the trick.
>
>
>
> 2017-10-23 15:05 GMT+03:00 Jim O'Flaherty :
>
>> Couldn't they be useful as part of a set of training data for newly
>> trained engines and networks?
>>
>> On Oct 23, 2017 2:34 AM, "Petri Pitkanen" 
>> wrote:
>>
>>> They are free to use in any attribution. Game score is a reflection of
>>> historical fact and hence not copyrightable. Dunno what use them are to
>>> anyone though.
>>>
>>> Petri
>>>
>>> 2017-10-23 2:29 GMT+03:00 Lucas Baker :
>>>
 Hi Robert,

 The AlphaGo Zero games are free to use with proper attribution, so
 please use them as you like for commentaries as long as you credit 
 DeepMind.

 Best,
 Lucas Baker

 On Sun, Oct 22, 2017 at 3:59 PM Robert Jasiek  wrote:

> AlphaGo Zero games are available as zipped SGF from Deepmind at
> http://www.alphago-games.com/ For earlier AlphaGo games, I have seen
> statements from Deepmind encouraging free use (presuming stating
> origin,
> of course) so that the games may be commented etc. I cannot find a
> similar statement from Deepmind for the published AlphaGo Zero games.
> Are they for free use or copyrighted? I hope the former so everybody
> including Deepmind can see more commentaries.
>
> --
> robert jasiek
> ___
> Computer-go mailing list
> Computer-go@computer-go.org
> http://computer-go.org/mailman/listinfo/computer-go


 ___
 Computer-go mailing list
 Computer-go@computer-go.org
 http://computer-go.org/mailman/listinfo/computer-go

>>>
>>>
>>> ___
>>> Computer-go mailing list
>>> Computer-go@computer-go.org
>>> http://computer-go.org/mailman/listinfo/computer-go
>>>
>>
>> ___
>> Computer-go mailing list
>> Computer-go@computer-go.org
>> http://computer-go.org/mailman/listinfo/computer-go
>>
>
>
> ___
> Computer-go mailing list
> Computer-go@computer-go.org
> http://computer-go.org/mailman/listinfo/computer-go
>
___
Computer-go mailing list
Computer-go@computer-go.org

Re: [Computer-go] AlphaGo Zero SGF - Free Use or Copyright?

[Xavier Combelle]

Hi Robert Jasiek,

you might have a delusional way to see the game of go and life, but I
would love that you would not pollute
my mailbox with such a delusional vision. I'm certain that a lot of
person of this mailing list and other forums share my view.

To sum up, I would be pleased and I'm quite certain others too that you
consider seriously behave more like others persons.

Did you already encounter a real game with "disturbing life kos or
anti-sekis" and especially "ladders (...) beyond 250 moves" ? If not how
do you believe that Alphago would learn how to manage such situations.

Xavier Combelle

Le 23/10/2017 à 16:35, Robert Jasiek a écrit :
> On 23.10.2017 14:05, Jim O'Flaherty wrote:
>> Couldn't they be useful as part of a set of training data for newly
>> trained
>> engines and networks?
>
> All the millions of games would be very useful for many purposes.
> E.g., I want to know whether the reconstructed knowledge includes such
> basic things as terminal positions with disturbing life kos or
> anti-sekis, whether ladders are recognised beyond 250 moves etc. Not
> to mention non-go applications.
>


___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] AlphaGo Zero SGF - Free Use or Copyright?

[Petri Pitkanen]

If the AG got better by playing against itself rather than training on
previous good players then I do not thing training data is that important.
Perhaps it is but google has shown that actually u dont need it. Just loads
of processing will do the trick.



2017-10-23 15:05 GMT+03:00 Jim O'Flaherty :

> Couldn't they be useful as part of a set of training data for newly
> trained engines and networks?
>
> On Oct 23, 2017 2:34 AM, "Petri Pitkanen" 
> wrote:
>
>> They are free to use in any attribution. Game score is a reflection of
>> historical fact and hence not copyrightable. Dunno what use them are to
>> anyone though.
>>
>> Petri
>>
>> 2017-10-23 2:29 GMT+03:00 Lucas Baker :
>>
>>> Hi Robert,
>>>
>>> The AlphaGo Zero games are free to use with proper attribution, so
>>> please use them as you like for commentaries as long as you credit DeepMind.
>>>
>>> Best,
>>> Lucas Baker
>>>
>>> On Sun, Oct 22, 2017 at 3:59 PM Robert Jasiek  wrote:
>>>
 AlphaGo Zero games are available as zipped SGF from Deepmind at
 http://www.alphago-games.com/ For earlier AlphaGo games, I have seen
 statements from Deepmind encouraging free use (presuming stating origin,
 of course) so that the games may be commented etc. I cannot find a
 similar statement from Deepmind for the published AlphaGo Zero games.
 Are they for free use or copyrighted? I hope the former so everybody
 including Deepmind can see more commentaries.

 --
 robert jasiek
 ___
 Computer-go mailing list
 Computer-go@computer-go.org
 http://computer-go.org/mailman/listinfo/computer-go
>>>
>>>
>>> ___
>>> Computer-go mailing list
>>> Computer-go@computer-go.org
>>> http://computer-go.org/mailman/listinfo/computer-go
>>>
>>
>>
>> ___
>> Computer-go mailing list
>> Computer-go@computer-go.org
>> http://computer-go.org/mailman/listinfo/computer-go
>>
>
> ___
> Computer-go mailing list
> Computer-go@computer-go.org
> http://computer-go.org/mailman/listinfo/computer-go
>
___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] AlphaGo Zero SGF - Free Use or Copyright?

[Robert Jasiek]

On 23.10.2017 14:05, Jim O'Flaherty wrote:

Couldn't they be useful as part of a set of training data for newly trained
engines and networks?


All the millions of games would be very useful for many purposes. E.g., 
I want to know whether the reconstructed knowledge includes such basic 
things as terminal positions with disturbing life kos or anti-sekis, 
whether ladders are recognised beyond 250 moves etc. Not to mention 
non-go applications.


--
robert jasiek
___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] AlphaGo Zero SGF - Free Use or Copyright?

[Álvaro Begué]

No, they are too few games for that.



On Mon, Oct 23, 2017 at 8:05 AM, Jim O'Flaherty 
wrote:

> Couldn't they be useful as part of a set of training data for newly
> trained engines and networks?
>
> On Oct 23, 2017 2:34 AM, "Petri Pitkanen" 
> wrote:
>
>> They are free to use in any attribution. Game score is a reflection of
>> historical fact and hence not copyrightable. Dunno what use them are to
>> anyone though.
>>
>> Petri
>>
>> 2017-10-23 2:29 GMT+03:00 Lucas Baker :
>>
>>> Hi Robert,
>>>
>>> The AlphaGo Zero games are free to use with proper attribution, so
>>> please use them as you like for commentaries as long as you credit DeepMind.
>>>
>>> Best,
>>> Lucas Baker
>>>
>>> On Sun, Oct 22, 2017 at 3:59 PM Robert Jasiek  wrote:
>>>
 AlphaGo Zero games are available as zipped SGF from Deepmind at
 http://www.alphago-games.com/ For earlier AlphaGo games, I have seen
 statements from Deepmind encouraging free use (presuming stating origin,
 of course) so that the games may be commented etc. I cannot find a
 similar statement from Deepmind for the published AlphaGo Zero games.
 Are they for free use or copyrighted? I hope the former so everybody
 including Deepmind can see more commentaries.

 --
 robert jasiek
 ___
 Computer-go mailing list
 Computer-go@computer-go.org
 http://computer-go.org/mailman/listinfo/computer-go
>>>
>>>
>>> ___
>>> Computer-go mailing list
>>> Computer-go@computer-go.org
>>> http://computer-go.org/mailman/listinfo/computer-go
>>>
>>
>>
>> ___
>> Computer-go mailing list
>> Computer-go@computer-go.org
>> http://computer-go.org/mailman/listinfo/computer-go
>>
>
> ___
> Computer-go mailing list
> Computer-go@computer-go.org
> http://computer-go.org/mailman/listinfo/computer-go
>
___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] AlphaGo Zero SGF - Free Use or Copyright?

[Jim O'Flaherty]

Couldn't they be useful as part of a set of training data for newly trained
engines and networks?

On Oct 23, 2017 2:34 AM, "Petri Pitkanen" 
wrote:

> They are free to use in any attribution. Game score is a reflection of
> historical fact and hence not copyrightable. Dunno what use them are to
> anyone though.
>
> Petri
>
> 2017-10-23 2:29 GMT+03:00 Lucas Baker :
>
>> Hi Robert,
>>
>> The AlphaGo Zero games are free to use with proper attribution, so please
>> use them as you like for commentaries as long as you credit DeepMind.
>>
>> Best,
>> Lucas Baker
>>
>> On Sun, Oct 22, 2017 at 3:59 PM Robert Jasiek  wrote:
>>
>>> AlphaGo Zero games are available as zipped SGF from Deepmind at
>>> http://www.alphago-games.com/ For earlier AlphaGo games, I have seen
>>> statements from Deepmind encouraging free use (presuming stating origin,
>>> of course) so that the games may be commented etc. I cannot find a
>>> similar statement from Deepmind for the published AlphaGo Zero games.
>>> Are they for free use or copyrighted? I hope the former so everybody
>>> including Deepmind can see more commentaries.
>>>
>>> --
>>> robert jasiek
>>> ___
>>> Computer-go mailing list
>>> Computer-go@computer-go.org
>>> http://computer-go.org/mailman/listinfo/computer-go
>>
>>
>> ___
>> Computer-go mailing list
>> Computer-go@computer-go.org
>> http://computer-go.org/mailman/listinfo/computer-go
>>
>
>
> ___
> Computer-go mailing list
> Computer-go@computer-go.org
> http://computer-go.org/mailman/listinfo/computer-go
>
___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] AlphaGo Zero SGF - Free Use or Copyright?

[Ingo Althöfer]

"Petri Pitkanen"  wrot:
>  They are free to use in any attribution. Game score is a reflection of 
> historical fact and hence not copyrightable. 


"reflection of historical fact" concerns games that were played
in public. Over the decades, there were several investigations
(and even Master theses) concerning copyright issues on chess
master games, played in public tournaments. In all cases (known to me) 
the conclusion was "they are free to use".

The case with private games (like in this case those of AG-0)
was not discussed.

Just my 2 Cent, Ingo.
___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] AlphaGo Zero SGF - Free Use or Copyright?

[Petri Pitkanen]

They are free to use in any attribution. Game score is a reflection of
historical fact and hence not copyrightable. Dunno what use them are to
anyone though.

Petri

2017-10-23 2:29 GMT+03:00 Lucas Baker :

> Hi Robert,
>
> The AlphaGo Zero games are free to use with proper attribution, so please
> use them as you like for commentaries as long as you credit DeepMind.
>
> Best,
> Lucas Baker
>
> On Sun, Oct 22, 2017 at 3:59 PM Robert Jasiek  wrote:
>
>> AlphaGo Zero games are available as zipped SGF from Deepmind at
>> http://www.alphago-games.com/ For earlier AlphaGo games, I have seen
>> statements from Deepmind encouraging free use (presuming stating origin,
>> of course) so that the games may be commented etc. I cannot find a
>> similar statement from Deepmind for the published AlphaGo Zero games.
>> Are they for free use or copyrighted? I hope the former so everybody
>> including Deepmind can see more commentaries.
>>
>> --
>> robert jasiek
>> ___
>> Computer-go mailing list
>> Computer-go@computer-go.org
>> http://computer-go.org/mailman/listinfo/computer-go
>
>
> ___
> Computer-go mailing list
> Computer-go@computer-go.org
> http://computer-go.org/mailman/listinfo/computer-go
>
___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] AlphaGo Zero SGF - Free Use or Copyright?

[Aja Huang]

2017-10-23 0:29 GMT+01:00 Lucas Baker :

> Hi Robert,
>
> The AlphaGo Zero games are free to use with proper attribution, so please
> use them as you like for commentaries as long as you credit DeepMind.
>

Yes, Lucas is right. We hope you enjoy AlphaGo Zero games. :)

Best regards,
Aja



> Best,
> Lucas Baker
>
> On Sun, Oct 22, 2017 at 3:59 PM Robert Jasiek  wrote:
>
>> AlphaGo Zero games are available as zipped SGF from Deepmind at
>> http://www.alphago-games.com/ For earlier AlphaGo games, I have seen
>> statements from Deepmind encouraging free use (presuming stating origin,
>> of course) so that the games may be commented etc. I cannot find a
>> similar statement from Deepmind for the published AlphaGo Zero games.
>> Are they for free use or copyrighted? I hope the former so everybody
>> including Deepmind can see more commentaries.
>>
>> --
>> robert jasiek
>> ___
>> Computer-go mailing list
>> Computer-go@computer-go.org
>> http://computer-go.org/mailman/listinfo/computer-go
>
>
> ___
> Computer-go mailing list
> Computer-go@computer-go.org
> http://computer-go.org/mailman/listinfo/computer-go
>
___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] AlphaGo Zero SGF - Free Use or Copyright?

[Lucas Baker]

Hi Robert,

The AlphaGo Zero games are free to use with proper attribution, so please
use them as you like for commentaries as long as you credit DeepMind.

Best,
Lucas Baker

On Sun, Oct 22, 2017 at 3:59 PM Robert Jasiek  wrote:

> AlphaGo Zero games are available as zipped SGF from Deepmind at
> http://www.alphago-games.com/ For earlier AlphaGo games, I have seen
> statements from Deepmind encouraging free use (presuming stating origin,
> of course) so that the games may be commented etc. I cannot find a
> similar statement from Deepmind for the published AlphaGo Zero games.
> Are they for free use or copyrighted? I hope the former so everybody
> including Deepmind can see more commentaries.
>
> --
> robert jasiek
> ___
> Computer-go mailing list
> Computer-go@computer-go.org
> http://computer-go.org/mailman/listinfo/computer-go
___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] AlphaGo Zero

[Gian-Carlo Pascutto]

On 21/10/2017 14:21, David Ongaro wrote:
> I understand that DeepMind might be unable to release the source code
> of AlphaGo due to policy or licensing reasons, but it would be great
> (and probably much more valuable) if they could release the fully
> trained network.

The source of AlphaGo Zero is really of zero interest (pun intended). It
can be obtained by ripping out ~50% of the Ray/Rn or AQ code (everything
related to MC playouts) and some minimal changes to evaluate the same
network for scoring and policy. Same for Leela.

It's literally possible to have a "Leela/Ray/AQ Zero" in a week or so
(it'll require a GPU or performance will be atrocious).

Of course, I can't give you the trained network to load into it. That'll
take another 88642 weeks.

So yes, the database of 29M self-play games would be immensely more
valuable. (Probably like the last 5M or so is fine, too). I prefer the
games over the network - with the games it's easier to train a smaller
network that gives better results on PC's that don't have 4 TPUs in them.

-- 
GCP
___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] AlphaGo Zero

[Hideki Kato]

The games look like previously published ones.  Just 
repeating?

Hideki

mic: <31fa8de6-c157-5de6-78fb-a66e6957a...@gmx.de>:
>There are several AlphaGo instances playing against each other on Tygem 

>at this moment.

>-Michael.

>

>On 21.10.2017 14:21, David Ongaro wrote:

>> Am 10/21/2017 um 03:12 AM schrieb uurtamo .:

>> 

>>> This sounds like a nice idea that is a misguided project.

>>>

>>> [...]

>>> Just accept that something awesome happened and that studying those 

>>> things that make it work well are more interesting than translating 

>>> coefficients into a bad understanding for people.

>>>

>>> I'm sorry that this NN can't teach anyone how to be a better player 

>>> through anything other than kicking their ass, but it wasn't built for 

>>> that.

>> 

>> Roberts approach might be misguided, but I don't agree that having the 

>> raw network data couldn't teach us something. E.g. have a look at this 

>> guy who was able to identify the neurons responsible for generating URLs 

>> in a wikipedia text generating RNN: 

>> 
>http://karpathy.github.io/2015/05/21/rnn-effectiveness/#visualizing-the-predictions-and-the-neuron-firings-in-the-rnn.

>> 

>> E.g. it might be possible to find the network Part of AlphaGo Zero which 

>> is responsible for L problems and use it to dream up new Problems! The 

>> possibilities could be endless. This kind of research might have been 

>> easier with the "classic" AlphaGo with separated policy and value 

>> networks, but should be possible anyways.

>> 

>> Also lets not forget DeepMinds own substantial research in this area: 

>> https://deepmind.com/blog/cognitive-psychology/.

>> 

>> I understand that DeepMind might be unable to release the source code of 

>> AlphaGo due to policy or licensing reasons, but it would be great (and 

>> probably much more valuable) if they could release the fully trained 

>> network. As Gian-Carlo Pascutto has pointed out, replicating this would 

>> not only incur high hardware costs but also take a long time.

>> 

>> David O.

>> 

>> 

>>> On Fri, Oct 20, 2017 at 8:24 AM, Robert Jasiek >> > wrote:

>>>

>>> On 20.10.2017 15:07, adrian.b.rob...@gmail.com

>>>  wrote:

>>>

>>> 1) Where is the semantic translation of the neural net to

>>> human theory

>>> knowledge?

>>>

>>> As far as (1), if we could do it, it would mean we could

>>> relate the

>>> structures embedded in the net's weight patterns to some other

>>> domain --

>>>

>>>

>>> The other domain can be "human go theory". It has various forms,

>>> from informal via textbook to mathematically proven. Sure, it is

>>> also incomplete but it can cope with additions.

>>>

>>> The neural net's weights and whatnot are given. This raw data can

>>> be deciphered in principle. By humans, algorithms or a combination.

>>>

>>> You do not know where to start? Why, that is easy: test! Modify

>>> ONE weight and study its effect on ONE aspect of human go theory,

>>> such as the occurrance (frequency) of independent life. No effect?

>>> Increase the modification, test a different weight, test a subset

>>> of adjacent weights etc. It has been possible to study semantics

>>> of parts of DNA, e.g., from differences related to illnesses.

>>> Modifications on the weights is like creating causes for illnesses

>>> (or improved health).

>>>

>>> There is no "we cannot do it", but maybe there is too much

>>> required effort for it to be financially worthwhile for the "too

>>> specialised" case of Go? As I say, a mathematical proof of a

>>> complete solution of Go will occur before AI playing perfectly;)

>>>

>>> So far neural

>>> nets have been trained and applied within single domains, and 
>any

>>> "generalization" means within that domain.

>>>

>>>

>>> Yes.

>>>

>>> -- 

>>> robert jasiek

>>>

>>> ___

>>> Computer-go mailing list

>>> Computer-go@computer-go.org 

>>> http://computer-go.org/mailman/listinfo/computer-go

>>> 

>>>

>>>

>>>

>>>

>>> ___

>>> Computer-go mailing list

>>> Computer-go@computer-go.org

>>> http://computer-go.org/mailman/listinfo/computer-go

>> 

>> 

>> 

>> ___

>> Computer-go mailing list

>> Computer-go@computer-go.org

>> http://computer-go.org/mailman/listinfo/computer-go

>> 

>

>---

>Diese E-Mail wurde von Avast Antivirus-Software auf Viren geprüft.

>https://www.avast.com/antivirus

>

>___

>Computer-go mailing list

>Computer-go@computer-go.org

Re: [Computer-go] AlphaGo Zero

[mic]

There are several AlphaGo instances playing against each other on Tygem 
at this moment.

-Michael.

On 21.10.2017 14:21, David Ongaro wrote:

Am 10/21/2017 um 03:12 AM schrieb uurtamo .:


This sounds like a nice idea that is a misguided project.

[...]
Just accept that something awesome happened and that studying those 
things that make it work well are more interesting than translating 
coefficients into a bad understanding for people.


I'm sorry that this NN can't teach anyone how to be a better player 
through anything other than kicking their ass, but it wasn't built for 
that.


Roberts approach might be misguided, but I don't agree that having the 
raw network data couldn't teach us something. E.g. have a look at this 
guy who was able to identify the neurons responsible for generating URLs 
in a wikipedia text generating RNN: 
http://karpathy.github.io/2015/05/21/rnn-effectiveness/#visualizing-the-predictions-and-the-neuron-firings-in-the-rnn.


E.g. it might be possible to find the network Part of AlphaGo Zero which 
is responsible for L problems and use it to dream up new Problems! The 
possibilities could be endless. This kind of research might have been 
easier with the "classic" AlphaGo with separated policy and value 
networks, but should be possible anyways.


Also lets not forget DeepMinds own substantial research in this area: 
https://deepmind.com/blog/cognitive-psychology/.


I understand that DeepMind might be unable to release the source code of 
AlphaGo due to policy or licensing reasons, but it would be great (and 
probably much more valuable) if they could release the fully trained 
network. As Gian-Carlo Pascutto has pointed out, replicating this would 
not only incur high hardware costs but also take a long time.


David O.


On Fri, Oct 20, 2017 at 8:24 AM, Robert Jasiek > wrote:


On 20.10.2017 15:07, adrian.b.rob...@gmail.com
 wrote:

1) Where is the semantic translation of the neural net to
human theory
knowledge?

As far as (1), if we could do it, it would mean we could
relate the
structures embedded in the net's weight patterns to some other
domain --


The other domain can be "human go theory". It has various forms,
from informal via textbook to mathematically proven. Sure, it is
also incomplete but it can cope with additions.

The neural net's weights and whatnot are given. This raw data can
be deciphered in principle. By humans, algorithms or a combination.

You do not know where to start? Why, that is easy: test! Modify
ONE weight and study its effect on ONE aspect of human go theory,
such as the occurrance (frequency) of independent life. No effect?
Increase the modification, test a different weight, test a subset
of adjacent weights etc. It has been possible to study semantics
of parts of DNA, e.g., from differences related to illnesses.
Modifications on the weights is like creating causes for illnesses
(or improved health).

There is no "we cannot do it", but maybe there is too much
required effort for it to be financially worthwhile for the "too
specialised" case of Go? As I say, a mathematical proof of a
complete solution of Go will occur before AI playing perfectly;)

So far neural
nets have been trained and applied within single domains, and any
"generalization" means within that domain.


Yes.

-- 
robert jasiek


___
Computer-go mailing list
Computer-go@computer-go.org 
http://computer-go.org/mailman/listinfo/computer-go





___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go




___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go



---
Diese E-Mail wurde von Avast Antivirus-Software auf Viren geprüft.
https://www.avast.com/antivirus

___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] AlphaGo Zero

[David Ongaro]

Am 10/21/2017 um 03:12 AM schrieb uurtamo .:


This sounds like a nice idea that is a misguided project.

[...]
Just accept that something awesome happened and that studying those 
things that make it work well are more interesting than translating 
coefficients into a bad understanding for people.


I'm sorry that this NN can't teach anyone how to be a better player 
through anything other than kicking their ass, but it wasn't built for 
that.


Roberts approach might be misguided, but I don't agree that having the 
raw network data couldn't teach us something. E.g. have a look at this 
guy who was able to identify the neurons responsible for generating URLs 
in a wikipedia text generating RNN: 
http://karpathy.github.io/2015/05/21/rnn-effectiveness/#visualizing-the-predictions-and-the-neuron-firings-in-the-rnn.


E.g. it might be possible to find the network Part of AlphaGo Zero which 
is responsible for L problems and use it to dream up new Problems! The 
possibilities could be endless. This kind of research might have been 
easier with the "classic" AlphaGo with separated policy and value 
networks, but should be possible anyways.


Also lets not forget DeepMinds own substantial research in this area: 
https://deepmind.com/blog/cognitive-psychology/.


I understand that DeepMind might be unable to release the source code of 
AlphaGo due to policy or licensing reasons, but it would be great (and 
probably much more valuable) if they could release the fully trained 
network. As Gian-Carlo Pascutto has pointed out, replicating this would 
not only incur high hardware costs but also take a long time.


David O.


On Fri, Oct 20, 2017 at 8:24 AM, Robert Jasiek > wrote:


On 20.10.2017 15:07, adrian.b.rob...@gmail.com
 wrote:

1) Where is the semantic translation of the neural net to
human theory
knowledge?

As far as (1), if we could do it, it would mean we could
relate the
structures embedded in the net's weight patterns to some other
domain --


The other domain can be "human go theory". It has various forms,
from informal via textbook to mathematically proven. Sure, it is
also incomplete but it can cope with additions.

The neural net's weights and whatnot are given. This raw data can
be deciphered in principle. By humans, algorithms or a combination.

You do not know where to start? Why, that is easy: test! Modify
ONE weight and study its effect on ONE aspect of human go theory,
such as the occurrance (frequency) of independent life. No effect?
Increase the modification, test a different weight, test a subset
of adjacent weights etc. It has been possible to study semantics
of parts of DNA, e.g., from differences related to illnesses.
Modifications on the weights is like creating causes for illnesses
(or improved health).

There is no "we cannot do it", but maybe there is too much
required effort for it to be financially worthwhile for the "too
specialised" case of Go? As I say, a mathematical proof of a
complete solution of Go will occur before AI playing perfectly;)

So far neural
nets have been trained and applied within single domains, and any
"generalization" means within that domain.


Yes.

-- 
robert jasiek


___
Computer-go mailing list
Computer-go@computer-go.org 
http://computer-go.org/mailman/listinfo/computer-go





___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go


___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] AlphaGo Zero

[Xavier Combelle]

Le 20/10/2017 à 17:24, Robert Jasiek a écrit :
>  Why, that is easy: test! Modify ONE weight and study its effect on
> ONE aspect of human go theory, such as the occurrance (frequency) of
> independent life. No effect? Increase the modification, test a
> different weight, test a subset of adjacent weights etc. It has been
> possible to study semantics of parts of DNA, e.g., from differences
> related to illnesses. Modifications on the weights is like creating
> causes for illnesses (or improved health).
I really believe you speak of things knowing nothing about. I'm far from
an expert in neural networks but from all that I know about them :

1- It really really unlikely to work
2- It is totally unrelated with DNA
3- It was probably already tried without success
___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] AlphaGo Zero

[Robert Jasiek]

On 20.10.2017 21:12, uurtamo . wrote:

do something like really careful experimental design across many dimensions
simultaneously (node weights) and several million experiments -- each of
which will require hundreds if not tens of thousands of games to find the
result of the change. Worse, there are probably tens of millions of neural
nets of this size that will perform equally well (isomorphisms plus minor
weight changes). So many changes will result in no change or a completely
useless game model.


It is possible that things turn out as complex as you describe...


"modeling through human knowledge" neural nets doesn't sound like a
sensible goal


...but I am not convinced. Researchers in the human brain's thinking 
keep their optimism, too.


Nevertheless, alternative approaches can be imagined. E.g., while 
building a neural net of eventually great strength also build in its own 
semantic interpretator, semantic verificator (including exclusion of 
errors as far as computationally possible) and translator between 
internal structure and human (or programming) language representation. I 
do not know if such dynamic self-representations of neural nets have 
already been described but if not this would be an interesting research 
topic.


--
robert jasiek
___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] AlphaGo Zero

[Gian-Carlo Pascutto]

On Fri, Oct 20, 2017, 21:48 Petr Baudis  wrote:

>   Few open questions I currently have, comments welcome:
>
>   - there is no input representing the number of captures; is this
> information somehow implicit or can the learned winrate predictor
> never truly approximate the true values because of this?
>

They are using Chinese rules, so prisoners don't matter. There are simply
less stones of one color on the board.


>   - what ballpark values for c_{puct} are reasonable?
>

The original paper has the value they used. But this likely needs tuning. I
would tune with a supervised network to get started, but you need games for
that. Does it even matter much early on? The network is random :)


>   - why is the dirichlet noise applied only at the root node, if it's
> useful?
>

It's only used to get some randomness in the move selection, no ? It's not
actually useful for anything besides that.


>   - the training process is quite lazy - it's not like the network sees
> each game immediately and adjusts, it looks at last 500k games and
> samples 1000*2048 positions, meaning about 4 positions per game (if
> I understood this right) - I wonder what would happen if we trained
> it more aggressively, and what AlphaGo does during the initial 500k
> games; currently, I'm training on all positions immediately, I guess
> I should at least shuffle them ;)
>

I think the lazyness may be related to the concern that reinforcement
methods can easily "forget" things they had learned before. The value
network training also likes positions from distinct games.


-- 

GCP
___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] AlphaGo Zero

[uurtamo .]

This sounds like a nice idea that is a misguided project.

Keep in mind the number of weights to change, and the fact that "one factor
at a time" testing will tell you nearly nothing about the overall dynamics
in a system of tens of thousands of dimensions. So you're going to need to
do something like really careful experimental design across many dimensions
simultaneously (node weights) and several million experiments -- each of
which will require hundreds if not tens of thousands of games to find the
result of the change. Worse, there are probably tens of millions of neural
nets of this size that will perform equally well (isomorphisms plus minor
weight changes). So many changes will result in no change or a completely
useless game model.

"modeling through human knowledge" neural nets doesn't sound like a
sensible goal -- it sounds more like a need to understand a topic in a
language not equipped for it without a simultaneous desire to understand a
topic under its own fundamental requirements in its own language.

Or you could build a machine-learning model to try to model those
changes except that you'd end up where you started, roughly. Another
black box and another frustrated human.

Just accept that something awesome happened and that studying those things
that make it work well are more interesting than translating coefficients
into a bad understanding for people.

I'm sorry that this NN can't teach anyone how to be a better player through
anything other than kicking their ass, but it wasn't built for that.

s.


On Fri, Oct 20, 2017 at 8:24 AM, Robert Jasiek  wrote:

> On 20.10.2017 15:07, adrian.b.rob...@gmail.com wrote:
>
>> 1) Where is the semantic translation of the neural net to human theory
>>> knowledge?
>>>
>> As far as (1), if we could do it, it would mean we could relate the
>> structures embedded in the net's weight patterns to some other domain --
>>
>
> The other domain can be "human go theory". It has various forms, from
> informal via textbook to mathematically proven. Sure, it is also incomplete
> but it can cope with additions.
>
> The neural net's weights and whatnot are given. This raw data can be
> deciphered in principle. By humans, algorithms or a combination.
>
> You do not know where to start? Why, that is easy: test! Modify ONE weight
> and study its effect on ONE aspect of human go theory, such as the
> occurrance (frequency) of independent life. No effect? Increase the
> modification, test a different weight, test a subset of adjacent weights
> etc. It has been possible to study semantics of parts of DNA, e.g., from
> differences related to illnesses. Modifications on the weights is like
> creating causes for illnesses (or improved health).
>
> There is no "we cannot do it", but maybe there is too much required effort
> for it to be financially worthwhile for the "too specialised" case of Go?
> As I say, a mathematical proof of a complete solution of Go will occur
> before AI playing perfectly;)
>
> So far neural
>> nets have been trained and applied within single domains, and any
>> "generalization" means within that domain.
>>
>
> Yes.
>
> --
> robert jasiek
>
> ___
> Computer-go mailing list
> Computer-go@computer-go.org
> http://computer-go.org/mailman/listinfo/computer-go
>
___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] AlphaGo Zero

[Gian-Carlo Pascutto]

On 19-10-17 13:00, Aja Huang via Computer-go wrote:
> Hi Hiroshi,
> 
> I think these are good questions. You can ask them at 
> https://www.reddit.com/r/MachineLearning/comments/76xjb5/ama_we_are_david_silver_and_julian_schrittwieser/

It seems the question was indeed asked but not answered:
https://www.reddit.com/r/MachineLearning/comments/76xjb5/ama_we_are_david_silver_and_julian_schrittwieser/dol03aq/

-- 
GCP
___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] AlphaGo Zero

[Robert Jasiek]

On 20.10.2017 15:07, adrian.b.rob...@gmail.com wrote:

1) Where is the semantic translation of the neural net to human theory
knowledge?

As far as (1), if we could do it, it would mean we could relate the
structures embedded in the net's weight patterns to some other domain --


The other domain can be "human go theory". It has various forms, from 
informal via textbook to mathematically proven. Sure, it is also 
incomplete but it can cope with additions.


The neural net's weights and whatnot are given. This raw data can be 
deciphered in principle. By humans, algorithms or a combination.


You do not know where to start? Why, that is easy: test! Modify ONE 
weight and study its effect on ONE aspect of human go theory, such as 
the occurrance (frequency) of independent life. No effect? Increase the 
modification, test a different weight, test a subset of adjacent weights 
etc. It has been possible to study semantics of parts of DNA, e.g., from 
differences related to illnesses. Modifications on the weights is like 
creating causes for illnesses (or improved health).


There is no "we cannot do it", but maybe there is too much required 
effort for it to be financially worthwhile for the "too specialised" 
case of Go? As I say, a mathematical proof of a complete solution of Go 
will occur before AI playing perfectly;)



So far neural
nets have been trained and applied within single domains, and any
"generalization" means within that domain.


Yes.

--
robert jasiek
___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] AlphaGo Zero

[Álvaro Begué]

When I did something like this for Spanish checkers (training a neural
network to be the evaluation function in an alpha-beta search, without any
human knowledge), I solved the problem of adding game variety by using UCT
for the opening moves. That means that I kept a tree structure with the
opening moves and I used the UCB1 formula to pick the next move as long as
the game was in the tree. Once outside the tree, I used alpha-beta search
to play a normal [very fast] game.

One important characteristic of this UCT opening-book builder is that the
last move inside the tree is basically random, so this explores a lot of
unbalanced positions.

Álvaro.



On Fri, Oct 20, 2017 at 9:23 AM, Petr Baudis  wrote:

>   I tried to reimplement the system - in a simplified way, trying to
> find the minimum that learns to play 5x5 in a few thousands of
> self-plays.  Turns out there are several components which are important
> to avoid some obvious attractors (like the network predicting black
> loses on every move from its second game on):
>
>   - disabling resignation in a portion of games is essential not just
> for tuning resignation threshold (if you want to even do that), but
> just to correct prediction signal by actual scoring rather than
> starting to always resign early in the game
>
>   - dirichlet (or other) noise is essential for the network getting
> looped into the same game - which is also self-reinforcing
>
>   - i have my doubts about the idea of high temperature move choices
> at the beginning, especially with T=1 ... maybe that's just bad
> very early in the training
>
> On Thu, Oct 19, 2017 at 02:23:41PM +0200, Petr Baudis wrote:
> >   The order of magnitude matches my parameter numbers.  (My attempt to
> > reproduce a simplified version of this is currently evolving at
> > https://github.com/pasky/michi/tree/nnet but the code is a mess right
> > now.)
>
> --
> Petr Baudis, Rossum
> Run before you walk! Fly before you crawl! Keep moving forward!
> If we fail, I'd rather fail really hugely.  -- Moist von Lipwig
> ___
> Computer-go mailing list
> Computer-go@computer-go.org
> http://computer-go.org/mailman/listinfo/computer-go
>
___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] AlphaGo Zero

[Gian-Carlo Pascutto]

On 19-10-17 13:23, Álvaro Begué wrote:
> Summing it all up, I get 22,837,864 parameters for the 20-block network
> and 46,461,544 parameters for the 40-block network.
> 
> Does this seem correct?

My Caffe model file is 185887898 bytes / 32-bit floats = 46 471 974

So yes, that seems pretty close. I'll send the model file and some
observations in a separate post.

-- 
GCP
___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] AlphaGo Zero

[Petr Baudis]

  I tried to reimplement the system - in a simplified way, trying to
find the minimum that learns to play 5x5 in a few thousands of
self-plays.  Turns out there are several components which are important
to avoid some obvious attractors (like the network predicting black
loses on every move from its second game on):

  - disabling resignation in a portion of games is essential not just
for tuning resignation threshold (if you want to even do that), but
just to correct prediction signal by actual scoring rather than
starting to always resign early in the game

  - dirichlet (or other) noise is essential for the network getting
looped into the same game - which is also self-reinforcing

  - i have my doubts about the idea of high temperature move choices
at the beginning, especially with T=1 ... maybe that's just bad
very early in the training

On Thu, Oct 19, 2017 at 02:23:41PM +0200, Petr Baudis wrote:
>   The order of magnitude matches my parameter numbers.  (My attempt to
> reproduce a simplified version of this is currently evolving at
> https://github.com/pasky/michi/tree/nnet but the code is a mess right
> now.)

-- 
Petr Baudis, Rossum
Run before you walk! Fly before you crawl! Keep moving forward!
If we fail, I'd rather fail really hugely.  -- Moist von Lipwig
___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] AlphaGo Zero

[Adrian . B . Robert]

Robert Jasiek  writes:

> So there is a superstrong neural net.
>
> 1) Where is the semantic translation of the neural net to human theory
> knowledge?
>
> 2) Where is the analysis of the neural net's errors in decision-making?
>
> 3) Where is the world-wide discussion preventing a combination of AI
> and (nano-)robots, which self-replicate or permanently ensure energy
> access, from causing extinction of mankind?


As far as (1), if we could do it, it would mean we could relate the
structures embedded in the net's weight patterns to some other domain --
if nothing else, the domain of the meanings of words in some natural
language.  We cannot, and most certainly the net cannot.  So far neural
nets have been trained and applied within single domains, and any
"generalization" means within that domain.  A net may learn to recognize
and act similarly with respect to a certain eye pattern on different
parts of the board.  No one, as far as I know, has presented a net that
would be able to use a guideline like, "two eyes alive, one eye dead" to
help it speed learning of how to act on the board.  But a human can
apply "one"/"two", and "alive"/"dead" once it has been made clear that
"eye" in this context is standing for a recognizable structure of
same-color-surrounding-space, and thereby learn in one step what the net
learns in thousands of incremental iterations.

And (2) presupposes (1), since to understand why a situation was
mis-perceived or mis-acted upon requires some understanding of what
exactly the perception and judgment process was in the first place.

It may be that the recent successes in brute-force learning powered by
improved hardware together with improved crafting of the architecture
eventually play some role in understanding and recreating
"intelligence".  But so far, using the term "AI" in connection with 99%
of this kind of work is just hype.  Useful in accomplishing engineering
goals, yes.  But not so much to do with intelligence.

___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] AlphaGo Zero

[Dan Schmidt]

On Fri, Oct 20, 2017 at 12:06 AM, Robert Jasiek  wrote:

>
> 3) Where is the world-wide discussion preventing a combination of AI and
> (nano-)robots, which self-replicate or permanently ensure energy access,
> from causing extinction of mankind?
>

You will find it if you Google for "artificial intelligence existential
threat". But the subject seems off-topic here.

Dan
___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] AlphaGo Zero

[Robert Jasiek]

On 20.10.2017 09:38, Xavier Combelle wrote:

What is currently named nanorobot is simply hand assembled molecules
which have mechanical properties and need huge
framework to be able simply move.


Sure. But we must not wait until such a thing exists.

--
robert jasiek
___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] AlphaGo Zero

[Xavier Combelle]

You seems to lack of knowing what is really a nano robot in current term.

They are very far to have the possibility to self replicate them self
and far more being able to dissolve the planet by doing that.

What is currently named nanorobot is simply hand assembled molecules
which have mechanical properties and need huge 

framework to be able simply move. So far to be a threat.


Le 20/10/2017 à 08:33, Robert Jasiek a écrit :
> On 20.10.2017 07:10, Petri Pitkanen wrote:
> >> 3) Where is the world-wide discussion preventing a combination of
> AI >> and (nano-)robots, which self-replicate or permanently ensure
> energy >> access, from causing extinction of mankind?
>> 3) Would it be a bad thing? All thing considered, not just human
>> point of
>> view
>
> Have you realised the potential of one successful self-duplication of
> a nano-robot? Iterate and the self-replicating nano-robots might
> dissolve the planet earth into elementary particles. Now discuss
> whether that might be good or bad. Not good for animals or plants, to
> start with.
>

___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] AlphaGo Zero

[Robert Jasiek]

On 20.10.2017 07:10, Petri Pitkanen wrote:
>> 3) Where is the world-wide discussion preventing a combination of AI 
>> and (nano-)robots, which self-replicate or permanently ensure energy 
>> access, from causing extinction of mankind?

3) Would it be a bad thing? All thing considered, not just human point of
view


Have you realised the potential of one successful self-duplication of a 
nano-robot? Iterate and the self-replicating nano-robots might dissolve 
the planet earth into elementary particles. Now discuss whether that 
might be good or bad. Not good for animals or plants, to start with.


--
robert jasiek
___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] AlphaGo Zero

[Petri Pitkanen]

1) There is no such thing and I do doubt if it ever will exist. Even humans
fail elaborate why they know certain things
2) If we are talking about new one. Very few people seen it playing so I
guess we lack the data. For the old we know it made errors, dunno if
analysis points why. Neural nets tend to be black boxes
3) Would it be a bad thing? All thing considered, not just human point of
view

2017-10-20 7:06 GMT+03:00 Robert Jasiek :

> So there is a superstrong neural net.
>
> 1) Where is the semantic translation of the neural net to human theory
> knowledge?
>
> 2) Where is the analysis of the neural net's errors in decision-making?
>
> 3) Where is the world-wide discussion preventing a combination of AI and
> (nano-)robots, which self-replicate or permanently ensure energy access,
> from causing extinction of mankind?
>
> --
> robert jasiek
>
> ___
> Computer-go mailing list
> Computer-go@computer-go.org
> http://computer-go.org/mailman/listinfo/computer-go
>
___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] AlphaGo Zero

[Robert Jasiek]

So there is a superstrong neural net.

1) Where is the semantic translation of the neural net to human theory 
knowledge?


2) Where is the analysis of the neural net's errors in decision-making?

3) Where is the world-wide discussion preventing a combination of AI and 
(nano-)robots, which self-replicate or permanently ensure energy access, 
from causing extinction of mankind?


--
robert jasiek
___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] AlphaGo Zero

[Álvaro Begué]

Yes, residual networks are awesome! I learned about them at ICML 2016 (
http://kaiminghe.com/icml16tutorial/index.html). Kaiming He's exposition
was fantastically clear. I used them in my own attempts at training neural
networks for move prediction. It's fairly easy to train something with 20
layers with residual networks, even without using batch normalization. With
batch normalization apparently you can get to hundreds of layers without
problems, and the models do perform better on the test data for vision
tasks. But I didn't implement that part, and the additional computational
cost probably makes this not worth it for go.

Álvaro.




On Thu, Oct 19, 2017 at 8:51 PM, Brian Sheppard via Computer-go <
computer-go@computer-go.org> wrote:

> So I am reading that residual networks are simply better than normal
> convolutional networks. There is a detailed write-up here:
> https://blog.waya.ai/deep-residual-learning-9610bb62c355
>
> Summary: the residual network has a fixed connection that adds (with no
> scaling) the output of the previous level to the output of the current
> level. The point is that once some layer learns a concept, that concept is
> immediately available to all downstream layers, without need for learning
> how to propagate the value through a complicated network design. These
> connections also provide a fast pathway for tuning deeper layers.
>
> -Original Message-
> From: Computer-go [mailto:computer-go-boun...@computer-go.org] On Behalf
> Of Gian-Carlo Pascutto
> Sent: Wednesday, October 18, 2017 4:33 PM
> To: computer-go@computer-go.org
> Subject: Re: [Computer-go] AlphaGo Zero
>
> On 18/10/2017 19:50, cazen...@ai.univ-paris8.fr wrote:
> >
> > https://deepmind.com/blog/
> >
> > http://www.nature.com/nature/index.html
>
> Select quotes that I find interesting from a brief skim:
>
> 1) Using a residual network was more accurate, achieved lower error, and
> improved performance in AlphaGo by over 600 Elo.
>
> 2) Combining policy and value together into a single network slightly
> reduced the move prediction accuracy, but reduced the value error and
> boosted playing performance in AlphaGo by around another 600 Elo.
>
> These gains sound very high (much higher than previous experiments with
> them reported here), but are likely due to the joint training.
>
> 3) The raw neural network, without using any lookahead, achieved an Elo
> rating of 3,055. ... AlphaGo Zero achieved a rating of 5,185.
>
> The increase of 2000 Elo from tree search sounds very high, but this may
> just mean the value network is simply very good - and perhaps relatively
> better than the policy one. (They previously had problems there that SL
> > RL for the policy network guiding the tree search - but I'm not sure
> there's any relation)
>
> 4) History features Xt; Yt are necessary because Go is not fully
> observable solely from the current stones, as repetitions are forbidden.
>
> This is a weird statement. Did they need 17 planes just to check for ko?
> It seems more likely that history features are very helpful for the
> internal understanding of the network as an optimization. That sucks though
> - it's annoying for analysis and position setup.
>
> Lastly, the entire training procedure is actually not very complicated at
> all, and it's hopeful the training is "faster" than previous approaches -
> but many things look fast if you can throw 64 GPU workers at a problem.
>
> In this context, the graphs of the differing network architectures causing
> huge strength discrepancies are both good and bad. Making a better pick can
> cause you to get massively better results, take a bad pick and you won't
> come close.
>
> --
> GCP
> ___
> Computer-go mailing list
> Computer-go@computer-go.org
> http://computer-go.org/mailman/listinfo/computer-go
>
> ___
> Computer-go mailing list
> Computer-go@computer-go.org
> http://computer-go.org/mailman/listinfo/computer-go
>
___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] AlphaGo Zero

[Brian Sheppard via Computer-go]

So I am reading that residual networks are simply better than normal 
convolutional networks. There is a detailed write-up here: 
https://blog.waya.ai/deep-residual-learning-9610bb62c355

Summary: the residual network has a fixed connection that adds (with no 
scaling) the output of the previous level to the output of the current level. 
The point is that once some layer learns a concept, that concept is immediately 
available to all downstream layers, without need for learning how to propagate 
the value through a complicated network design. These connections also provide 
a fast pathway for tuning deeper layers.

-Original Message-
From: Computer-go [mailto:computer-go-boun...@computer-go.org] On Behalf Of 
Gian-Carlo Pascutto
Sent: Wednesday, October 18, 2017 4:33 PM
To: computer-go@computer-go.org
Subject: Re: [Computer-go] AlphaGo Zero

On 18/10/2017 19:50, cazen...@ai.univ-paris8.fr wrote:
> 
> https://deepmind.com/blog/
> 
> http://www.nature.com/nature/index.html

Select quotes that I find interesting from a brief skim:

1) Using a residual network was more accurate, achieved lower error, and 
improved performance in AlphaGo by over 600 Elo.

2) Combining policy and value together into a single network slightly reduced 
the move prediction accuracy, but reduced the value error and boosted playing 
performance in AlphaGo by around another 600 Elo.

These gains sound very high (much higher than previous experiments with them 
reported here), but are likely due to the joint training.

3) The raw neural network, without using any lookahead, achieved an Elo rating 
of 3,055. ... AlphaGo Zero achieved a rating of 5,185.

The increase of 2000 Elo from tree search sounds very high, but this may just 
mean the value network is simply very good - and perhaps relatively better than 
the policy one. (They previously had problems there that SL
> RL for the policy network guiding the tree search - but I'm not sure
there's any relation)

4) History features Xt; Yt are necessary because Go is not fully observable 
solely from the current stones, as repetitions are forbidden.

This is a weird statement. Did they need 17 planes just to check for ko?
It seems more likely that history features are very helpful for the internal 
understanding of the network as an optimization. That sucks though - it's 
annoying for analysis and position setup.

Lastly, the entire training procedure is actually not very complicated at all, 
and it's hopeful the training is "faster" than previous approaches - but many 
things look fast if you can throw 64 GPU workers at a problem.

In this context, the graphs of the differing network architectures causing huge 
strength discrepancies are both good and bad. Making a better pick can cause 
you to get massively better results, take a bad pick and you won't come close.

--
GCP
___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] AlphaGo Zero

[dave.de...@planet.nl]

I would like to know how much handicap the Master version needs against the 
Zero version. It could be less than black without komi or more than 3 stones.
Handicap differences cannot be deduced from regular Elo rating differences, 
because it varies depending on skill (a handicap stone is more than 100 regular 
Elo points points for higher dan players and less than 100 regular Elo points 
for kyu players). 
Dave de Vos
 
>Origineel Bericht--
--
>Van : 3-hirn-ver...@gmx.de
>Datum : 19/10/2017 20:53
>Aan : computer-go@computer-go.org
>Onderwerp : Re: [Computer-go] AlphaGo Zero
>
>What shall I say?
>Really impressive.
>My congratulations to the DeepMind team!
>
>> https://deepmind.com/blog/
>> http://www.nature.com/nature/index.html
>
>* Would the same approach also work for integral komi values
>(with the possibility of draws)? If so, what would the likely
>correct komi for 19x19 Go be?
>
>* Or in another way: Looking at Go on NxN board:
>For which values of N would the DeepMind be confident
>to find the correct komi value?
>
>
>* How often are there ko-fights in autoplay games of
>AlphaGo Zero?
>
>Ingo.
>
>PS(a fitting song). The opening theme of
>Djan-Go Unchained (with a march through a desert of stones):
>https://www.youtube.com/watch?v=R1hqn8kKZ_M
>___
>Computer-go mailing list
>Computer-go@computer-go.org
>http://computer-go.org/mailman/listinfo/computer-go
___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] AlphaGo Zero

[Ingo Althöfer]

What shall I say?
Really impressive.
My congratulations to the DeepMind team!

> https://deepmind.com/blog/
> http://www.nature.com/nature/index.html

* Would the same approach also work for integral komi values
(with the possibility of draws)? If so, what would the likely
correct komi for 19x19 Go be?

* Or in another way: Looking at Go on NxN board:
For which values of N would the DeepMind be confident
to find the correct komi value?


* How often are there ko-fights in autoplay games of
AlphaGo Zero?

Ingo.

PS(a fitting song). The opening theme of
Djan-Go Unchained (with a march through a desert of stones):
https://www.youtube.com/watch?v=R1hqn8kKZ_M
___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] AlphaGo Zero

[Álvaro Begué]

Yes, it seems really odd that they didn't add a plane of all ones. The
"heads" have weights that depend on the location of the board, but all the
other layers can't tell the difference between a lonely stone at (1,1) and
one at (3,3).

In my own experiments (trying to predict human moves) I found that 3 inputs
worked well: signed liberties, age capped at 8, all ones. I think of the
number of liberties as a key part of the game mechanics, so I don't think
it detracts from the purity of the approach, and it's probably helpful for
learning about life and death.

Álvaro.




On Thu, Oct 19, 2017 at 7:42 AM, Gian-Carlo Pascutto  wrote:

> On 18-10-17 19:50, cazen...@ai.univ-paris8.fr wrote:
> >
> > https://deepmind.com/blog/
> >
> > http://www.nature.com/nature/index.html
>
> Another interesting tidbit:
>
> The inputs don't contain a reliable board edge. The "white to move"
> plane contains it, but only when white is to move.
>
> So until AG Zero "black" learned that a go board is 19 x 19, the white
> player had a serious advantage.
>
> I think I will use 18 input layers :-)
>
> --
> GCP
> ___
> Computer-go mailing list
> Computer-go@computer-go.org
> http://computer-go.org/mailman/listinfo/computer-go
>
___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] AlphaGo Zero

[Petr Baudis]

  The order of magnitude matches my parameter numbers.  (My attempt to
reproduce a simplified version of this is currently evolving at
https://github.com/pasky/michi/tree/nnet but the code is a mess right
now.)

On Thu, Oct 19, 2017 at 07:23:31AM -0400, Álvaro Begué wrote:
> This is a quick check of my understanding of the network architecture.
> Let's count the number of parameters in the model:
>  * convolutional block: (17*9+1)*256 + 2*256
> [ 17 = number of input channels
>9 = size of the 3x3 convolution window
>1 = bias (I am not sure this is needed if you are going to do batch
> normalization immediately after)
>  256 = number of output channels
>2 = mean and standard deviation of the output of the batch normalization
>  256 = number of channels in the batch normalization ]
>  * residual block: (256*9+1)*256 + 2*256 + (256*9+1)*256 + 2*256
>  * policy head: (256*1+1)*2 + 2*2 + (2*361+1)*362
>  * value head: (256*1+1)*1 + 2*1 + (1*361+1)*256 + (256+1)*1
> 
> Summing it all up, I get 22,837,864 parameters for the 20-block network and
> 46,461,544 parameters for the 40-block network.
> 
> Does this seem correct?
> 
> Álvaro.
> 
> 
> 
> On Thu, Oct 19, 2017 at 6:17 AM, Petr Baudis  wrote:
> 
> > On Wed, Oct 18, 2017 at 04:29:47PM -0700, David Doshay wrote:
> > > I saw my first AlphaGo Zero joke today:
> > >
> > > After a few more months of self-play the games might look like this:
> > >
> > > AlphaGo Zero Black - move 1
> > > AlphaGo Zero White - resigns
> >
> > ...which is exactly what my quick attempt to reproduce AlphaGo Zero
> > yesterday converged to overnight. ;-)  But I'm afraid it's because of
> > a bug, not wisdom...
> >
> > Petr Baudis
> > ___
> > Computer-go mailing list
> > Computer-go@computer-go.org
> > http://computer-go.org/mailman/listinfo/computer-go
> >

> ___
> Computer-go mailing list
> Computer-go@computer-go.org
> http://computer-go.org/mailman/listinfo/computer-go


-- 
Petr Baudis, Rossum
Run before you walk! Fly before you crawl! Keep moving forward!
If we fail, I'd rather fail really hugely.  -- Moist von Lipwig
___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] AlphaGo Zero

[Gian-Carlo Pascutto]

On 18-10-17 19:50, cazen...@ai.univ-paris8.fr wrote:
> 
> https://deepmind.com/blog/
> 
> http://www.nature.com/nature/index.html

Another interesting tidbit:

The inputs don't contain a reliable board edge. The "white to move"
plane contains it, but only when white is to move.

So until AG Zero "black" learned that a go board is 19 x 19, the white
player had a serious advantage.

I think I will use 18 input layers :-)

-- 
GCP
___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] AlphaGo Zero

[Álvaro Begué]

This is a quick check of my understanding of the network architecture.
Let's count the number of parameters in the model:
 * convolutional block: (17*9+1)*256 + 2*256
[ 17 = number of input channels
   9 = size of the 3x3 convolution window
   1 = bias (I am not sure this is needed if you are going to do batch
normalization immediately after)
 256 = number of output channels
   2 = mean and standard deviation of the output of the batch normalization
 256 = number of channels in the batch normalization ]
 * residual block: (256*9+1)*256 + 2*256 + (256*9+1)*256 + 2*256
 * policy head: (256*1+1)*2 + 2*2 + (2*361+1)*362
 * value head: (256*1+1)*1 + 2*1 + (1*361+1)*256 + (256+1)*1

Summing it all up, I get 22,837,864 parameters for the 20-block network and
46,461,544 parameters for the 40-block network.

Does this seem correct?

Álvaro.



On Thu, Oct 19, 2017 at 6:17 AM, Petr Baudis  wrote:

> On Wed, Oct 18, 2017 at 04:29:47PM -0700, David Doshay wrote:
> > I saw my first AlphaGo Zero joke today:
> >
> > After a few more months of self-play the games might look like this:
> >
> > AlphaGo Zero Black - move 1
> > AlphaGo Zero White - resigns
>
> ...which is exactly what my quick attempt to reproduce AlphaGo Zero
> yesterday converged to overnight. ;-)  But I'm afraid it's because of
> a bug, not wisdom...
>
> Petr Baudis
> ___
> Computer-go mailing list
> Computer-go@computer-go.org
> http://computer-go.org/mailman/listinfo/computer-go
>
___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] AlphaGo Zero

[Aja Huang via Computer-go]

On Thu, Oct 19, 2017 at 11:04 AM, Hiroshi Yamashita 
wrote:

> I have two questions.
>
> 2017 Jan, Master , defeat 60 pros in a row.
> 2017 May, Master?, defeat Ke Jie 3-0.
>
> Master is Zero method with rollout.
> Zero   is Zero method without rollout.
>
> Did AlphaGo that played with Ke Jie use rollout?
> Is Zero with rollout stronger than Zero without rollout?
>

Hi Hiroshi,

I think these are good questions. You can ask them at
https://www.reddit.com/r/MachineLearning/comments/76xjb5/ama_we_are_david_silver_and_julian_schrittwieser/

Aja


> Thanks,
> Hiroshi Yamashita
>
> - Original Message - From: 
> To: 
> Sent: Thursday, October 19, 2017 2:50 AM
> Subject: [Computer-go] AlphaGo Zero
>
>
>
>> https://deepmind.com/blog/
>>
>> http://www.nature.com/nature/index.html
>>
>> Impressive!
>>
>
> ___
> Computer-go mailing list
> Computer-go@computer-go.org
> http://computer-go.org/mailman/listinfo/computer-go
>
___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] AlphaGo Zero

[Petr Baudis]

On Wed, Oct 18, 2017 at 04:29:47PM -0700, David Doshay wrote:
> I saw my first AlphaGo Zero joke today:
> 
> After a few more months of self-play the games might look like this:
> 
> AlphaGo Zero Black - move 1
> AlphaGo Zero White - resigns

...which is exactly what my quick attempt to reproduce AlphaGo Zero
yesterday converged to overnight. ;-)  But I'm afraid it's because of
a bug, not wisdom...

Petr Baudis
___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] AlphaGo Zero

[Hiroshi Yamashita]

I have two questions.

2017 Jan, Master , defeat 60 pros in a row.
2017 May, Master?, defeat Ke Jie 3-0.

Master is Zero method with rollout.
Zero   is Zero method without rollout.

Did AlphaGo that played with Ke Jie use rollout?
Is Zero with rollout stronger than Zero without rollout?

Thanks,
Hiroshi Yamashita

- Original Message - 
From: 

To: 
Sent: Thursday, October 19, 2017 2:50 AM
Subject: [Computer-go] AlphaGo Zero



https://deepmind.com/blog/

http://www.nature.com/nature/index.html

Impressive!


___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go

Re: [Computer-go] AlphaGo Zero

[Joona Kiiski]

About the fact that ladders appeared so late:

- The learning was based on self-play. Understanding ladders is perhaps not
so important if your opponent doesn't understand them either... Every time
a decisive ladder appears on the board, the result is practically a coin
toss.

- And as others have pointed out, unlike almost all other go features,
ladders are not at all a local feature. The features need to build up
through a huge number of convolution layers, before it works. And it's
difficult to build this understanding incrementally (unlike e.g. life &
death where you can start with simple cases and then move to more difficult
cases), so we lack bias to direct the learning to the right direction.

On Wed, Oct 18, 2017 at 3:04 PM, Brian Sheppard via Computer-go <
computer-go@computer-go.org> wrote:

> Some thoughts toward the idea of general game-playing...
>
> One aspect of Go is ideally suited for visual NN: strong locality of
> reference.  That is, stones affect stones that are nearby.
>
> I wonder whether the late emergence of ladder understanding within AlphaGo
> Zero is an artifact of the board representation? The authors speculate that
> ladders are not as important as humans surmise.
>
> Another aspect of Go is ideally suited for visual NN: translation
> invariance. The convolutions transfer knowledge around the board, with the
> presumption that good moves will travel well.
>
> I wonder whether we can data-mine positional evaluations to discover
> features? E.g., start with a standard visual NN and then make a database of
> positions where the delta between actual and expected evaluation is large
> enough to cause a different move to be selected. Data mine features from
> this set, and extend the NN with new inputs. (This would not discover a
> notion like "liberty count", but would discover notions like "ladder
> breaker".)
>
>
> -Original Message-
> From: Computer-go [mailto:computer-go-boun...@computer-go.org] On Behalf
> Of Gian-Carlo Pascutto
> Sent: Wednesday, October 18, 2017 4:33 PM
> To: computer-go@computer-go.org
> Subject: Re: [Computer-go] AlphaGo Zero
>
> On 18/10/2017 19:50, cazen...@ai.univ-paris8.fr wrote:
> >
> > https://deepmind.com/blog/
> >
> > http://www.nature.com/nature/index.html
>
> Select quotes that I find interesting from a brief skim:
>
> 1) Using a residual network was more accurate, achieved lower error, and
> improved performance in AlphaGo by over 600 Elo.
>
> 2) Combining policy and value together into a single network slightly
> reduced the move prediction accuracy, but reduced the value error and
> boosted playing performance in AlphaGo by around another 600 Elo.
>
> These gains sound very high (much higher than previous experiments with
> them reported here), but are likely due to the joint training.
>
> 3) The raw neural network, without using any lookahead, achieved an Elo
> rating of 3,055. ... AlphaGo Zero achieved a rating of 5,185.
>
> The increase of 2000 Elo from tree search sounds very high, but this may
> just mean the value network is simply very good - and perhaps relatively
> better than the policy one. (They previously had problems there that SL
> > RL for the policy network guiding the tree search - but I'm not sure
> there's any relation)
>
> 4) History features Xt; Yt are necessary because Go is not fully
> observable solely from the current stones, as repetitions are forbidden.
>
> This is a weird statement. Did they need 17 planes just to check for ko?
> It seems more likely that history features are very helpful for the
> internal understanding of the network as an optimization. That sucks though
> - it's annoying for analysis and position setup.
>
> Lastly, the entire training procedure is actually not very complicated at
> all, and it's hopeful the training is "faster" than previous approaches -
> but many things look fast if you can throw 64 GPU workers at a problem.
>
> In this context, the graphs of the differing network architectures causing
> huge strength discrepancies are both good and bad. Making a better pick can
> cause you to get massively better results, take a bad pick and you won't
> come close.
>
> --
> GCP
> ___
> Computer-go mailing list
> Computer-go@computer-go.org
> http://computer-go.org/mailman/listinfo/computer-go
>
> ___
> Computer-go mailing list
> Computer-go@computer-go.org
> http://computer-go.org/mailman/listinfo/computer-go
>
___
Computer-go mailing list
Computer-go@computer-go.org
http://computer-go.org/mailman/listinfo/computer-go