Re: [Computer-go] Neural networks + MCTS applied to chemical syntheses

[Brian Lee]

In particular, they had no way to train a value net, so it was back to
AlphaGo v1 style of training just a policy net and reusing it as the
rollout policy.



On Fri, Apr 6, 2018 at 6:31 AM Fidel Santiago  wrote:

> Hello,
>
> Apparently the lessons of Alphago (and many others) are being applied to
> other fields:
>
> https://www.nature.com/articles/d41586-018-03774-5
>
> "The authors devised a computational process that starts by automatically
> extracting chemical transformations from a large commercial database, being
> careful to include only reactions that have been reported several times.
> Their system accepts these well-precedented reactions as ‘allowed moves’ in
> organic synthesis. When the system is asked to devise a synthetic route to
> a target molecule, it works backwards from the target as would a human,
> picking out the most promising precursor molecules according to the design
> rules that it has learnt, and then seeing how feasible it is to synthesize
> those. The authors combined three artificial neural networks with a random
> Monte Carlo tree search — a type of search algorithm used by computers in
> certain decision-making processes — to narrow down the most promising
> synthetic routes, without getting stuck too quickly on a particular path."
>
> Ciao!
>
> Fidel Santiago.
> ___
> 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] Neural networks + MCTS applied to chemical syntheses

[Fidel Santiago]

Hello,

Apparently the lessons of Alphago (and many others) are being applied to
other fields:

https://www.nature.com/articles/d41586-018-03774-5

"The authors devised a computational process that starts by automatically
extracting chemical transformations from a large commercial database, being
careful to include only reactions that have been reported several times.
Their system accepts these well-precedented reactions as ‘allowed moves’ in
organic synthesis. When the system is asked to devise a synthetic route to
a target molecule, it works backwards from the target as would a human,
picking out the most promising precursor molecules according to the design
rules that it has learnt, and then seeing how feasible it is to synthesize
those. The authors combined three artificial neural networks with a random
Monte Carlo tree search — a type of search algorithm used by computers in
certain decision-making processes — to narrow down the most promising
synthetic routes, without getting stuck too quickly on a particular path."

Ciao!

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

Re: [computer-go] Neural networks

[steve uurtamo]

these things have definitions, folks.

everything isn't everything else.

s.

On Tue, Nov 3, 2009 at 7:43 AM, Willemien wilem...@googlemail.com wrote:
 I disagree with the point that MCTS is a neural network,

 In my opinion (and i maybe completely off target) One of the essences
 of neural networks is that the program changes/learns from the games
 it has played. .

 MCTS doesn't have that result, the improvement is only in-game
 The program doesn't learn not to make the same mistake anymore, by
 MCTS the mistake is hopefully avoided.

 if MTCS is a neyural network than alpha beta would also qualify.

 but i may be wrong it is just an opening for a discussion in which we
 all can learn


 2009/11/2 Daniel Burgos dbur...@gmail.com:
 Well, at its esence a computer is an universal Turing Machine. If you
 organize the program as a neural network or as a MC algorithm is just
 cosmetics.

 You can see the circuits of your computer as neurons simulating a Von
 Neumann architecture that is simulating a neural network or a MC or
 whatever. An may be your whatever is simulating another thing.

 2009/11/2 Álvaro Begué alvaro.be...@gmail.com

 On Sun, Nov 1, 2009 at 7:50 PM, Aldric Giacomoni ald...@trevoke.net
 wrote:
  Álvaro Begué wrote:
  2009/10/31  compgo...@aol.com:
  Present day MC Go programs are neural networks. The playout is the
  trainng
  process.
 
  What?
  ___
  computer-go mailing list
  computer-go@computer-go.org
  http://www.computer-go.org/mailman/listinfo/computer-go/
 
  Go programs using the Monte-Carlo algorithms are neural networks. The
  actual
  fact of playing games is what trains them to play better.
 
  ... I'm pretty sure that's what it means, anyway.

 Except they don't fit any definition of neural network that I've
 been able to find. For starters, they don't have neurons.
 ___
 computer-go mailing list
 computer-go@computer-go.org
 http://www.computer-go.org/mailman/listinfo/computer-go/


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

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

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

Re: [computer-go] Neural networks

[Richard Brown]

On Tue, Nov 3, 2009 at 6:43 AM, Willemien wilem...@googlemail.com wrote:
 I disagree with the point that MCTS is a neural network,

 In my opinion (and i maybe completely off target) One of the essences
 of neural networks is that the program changes/learns from the games
 it has played. .

I think that you are right; the learning associated with artificial neural nets
is likely to be from games the program has played, or from games it has
observed.  In some literature, this is called supervised learning, or
learning with a teacher, wherein the training data is not at all random,
but comprises loads of very specific examples of actual behavior on the
board.  (Of course, deciding what features of those actual behaviors are
to be measured, or encoded into a pattern, is a very tricky problem.)

 MCTS doesn't have that result, the improvement is only in-game
 The program doesn't learn not to make the same mistake anymore, by
 MCTS the mistake is hopefully avoided.

MCTS seems to be (I'm sure someone will correct me if I'm wrong)
reinforcement learning which differs from supervised learning
in that correct input/output pairs are never explicitly shown to the
classifier.

Just tons of nearly-random trials, yes?

According to  http://en.wikipedia.org/wiki/Reinforcement_learning ,  there
is a focus on on-line performance, which involves finding a balance between
exploration (of uncharted territory) and exploitation (of current knowledge).
The exploration vs. exploitation trade-off in reinforcement learning has been
mostly studied through the multi-armed bandit problem.

So, the two methods are both classifiers of data points, in some sense.

The fundamental difference between them, IMHO, for the purposes of
go-programming, is that artificial neural nets may learn from data that is
harvested (and preserved, at least for the duration of the training), while
Monte-Carlo methods learn from (mostly random) unsupervised trial-and-error.

Further, the tuning of the weights in an artificial neural net may be performed
off-line, with data from thousands of games, while MCTS performs its work
on-line, with a branch of the game-tree that begins only from the
current position.

Are both methods classifiers?  Sure.  But neural net training is done through
observation/measurement/feature-extraction (offline, from thousands of games).
That's a much different critter, at the core, from the environment/action/reward
scenario of MCTS (online, from the current position).

-- 
All computer programs are identical.  After all, it's just ones and zeroes,
and a paper tape of arbitrary length.
___
computer-go mailing list
computer-go@computer-go.org
http://www.computer-go.org/mailman/listinfo/computer-go/

Re: [computer-go] Neural networks

[Daniel Burgos]

Well, at its esence a computer is an universal Turing Machine. If you
organize the program as a neural network or as a MC algorithm is just
cosmetics.

You can see the circuits of your computer as neurons simulating a Von
Neumann architecture that is simulating a neural network or a MC or
whatever. An may be your whatever is simulating another thing.

2009/11/2 Álvaro Begué alvaro.be...@gmail.com

 On Sun, Nov 1, 2009 at 7:50 PM, Aldric Giacomoni ald...@trevoke.net
 wrote:
  Álvaro Begué wrote:
  2009/10/31  compgo...@aol.com:
  Present day MC Go programs are neural networks. The playout is the
 trainng
  process.
 
  What?
  ___
  computer-go mailing list
  computer-go@computer-go.org
  http://www.computer-go.org/mailman/listinfo/computer-go/
 
  Go programs using the Monte-Carlo algorithms are neural networks. The
 actual
  fact of playing games is what trains them to play better.
 
  ... I'm pretty sure that's what it means, anyway.

 Except they don't fit any definition of neural network that I've
 been able to find. For starters, they don't have neurons.
 ___
 computer-go mailing list
 computer-go@computer-go.org
 http://www.computer-go.org/mailman/listinfo/computer-go/

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

Re: [computer-go] Neural networks

[Álvaro Begué]

2009/10/31  compgo...@aol.com:
 Present day MC Go programs are neural networks. The playout is the trainng
 process.

What?
___
computer-go mailing list
computer-go@computer-go.org
http://www.computer-go.org/mailman/listinfo/computer-go/

Re: [computer-go] Neural networks

[Aldric Giacomoni]

Álvaro Begué wrote:
 2009/10/31  compgo...@aol.com:
 Present day MC Go programs are neural networks. The playout is the trainng
 process.
 
 What?
 ___
 computer-go mailing list
 computer-go@computer-go.org
 http://www.computer-go.org/mailman/listinfo/computer-go/

Go programs using the Monte-Carlo algorithms are neural networks. The actual
fact of playing games is what trains them to play better.

... I'm pretty sure that's what it means, anyway.

-- 
Aldric Giacomoni
Truth has no special time of its own.  Its hour is now -- always.
-- Albert Schweitzer
___
computer-go mailing list
computer-go@computer-go.org
http://www.computer-go.org/mailman/listinfo/computer-go/

Re: [computer-go] Neural networks

[Álvaro Begué]

On Sun, Nov 1, 2009 at 7:50 PM, Aldric Giacomoni ald...@trevoke.net wrote:
 Álvaro Begué wrote:
 2009/10/31  compgo...@aol.com:
 Present day MC Go programs are neural networks. The playout is the trainng
 process.

 What?
 ___
 computer-go mailing list
 computer-go@computer-go.org
 http://www.computer-go.org/mailman/listinfo/computer-go/

 Go programs using the Monte-Carlo algorithms are neural networks. The actual
 fact of playing games is what trains them to play better.

 ... I'm pretty sure that's what it means, anyway.

Except they don't fit any definition of neural network that I've
been able to find. For starters, they don't have neurons.
___
computer-go mailing list
computer-go@computer-go.org
http://www.computer-go.org/mailman/listinfo/computer-go/

Re: [computer-go] Neural networks

[compgo123]

Present day MC Go programs are neural networks. The playout is the trainng 
process.

DL


-Original Message-
From: Petr Baudis pa...@ucw.cz
To: computer-go@computer-go.org
Sent: Wed, Oct 14, 2009 7:26 am
Subject: [computer-go] Neural networks



  Hi!
  Is there some high-level reason hypothesised about why there are
o successful programs using neural networks in Go?
  I'd also like to ask if someone has a research tip for some
nteresting Go sub-problem that could make for a nice beginner neural
etworks project.
  Thanks,
-- 
   Petr Pasky Baudis
 lot of people have my books on their bookshelves.
hat's the problem, they need to read them. -- Don Knuth
__
omputer-go mailing list
omputer...@computer-go.org
ttp://www.computer-go.org/mailman/listinfo/computer-go/

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

Re: [computer-go] Neural networks

[Petr Baudis]

On Wed, Oct 14, 2009 at 05:12:58PM +0200, Heikki Levanto wrote:
 On Wed, Oct 14, 2009 at 03:34:59PM +0300, Petri Pitkanen wrote:
  Neural network tend to work well in those cases where evaluation function is
  smooth, like backgammon. Even inbackgammon neural networks do give good
  results if situation has possibility of sudden equity changes like deep
  backgames and deep anchor games. Top backgammon programs 3-ply search on top
  neural network to handle these problems.
  
  I do not know wher neural nets would fit well, perhaps finding invasion
  spots?
 
 I have been speculating about a NN evaluation function for go, feeding it a
 lot of preprocessed information about the position, like number of strings
 with 1,2,3,4, or more liberties, number of stones in same, number of separate
 groups, number of obviously dead stones, strings, and groups, number of
 points clearly controlled by each player, etc, etc. This should be possible
 to train from existing games where we know the result (in the beginning it is
 50-50, in the end one or the other has won 100-0. Assume some simple function
 in between). 

Interesting idea, but I think MCTS programs already handle quite fine
the part where there is some simple function in between. Most of the
problematic games seem to be when MCTS is chasing and trying to kill a
single group, and the question is whether it succeeds; IOW, I think the
problem is tactical rather than strategical in MCTS (I have few of my
own ideas to try to deal with that, non-related to NN ;-).

I like Magnus' ideas, though the 3x3 problem seems to be a bit of a
special case of what Remi did, and I don't really have an idea how to
well train the in-UCT NN.

Another idea I received was actually not related to computers _playing_
Go, but maybe that's all the more interesting in current times - a NN
that could identify the player's playing style by looking at few of
her games and deciding which well-known player they look like (e.g.
Takemiya, Cho Chikun, Lee Sedol, TheCaptain ;-). Of course it is quite
an interesting problem what set of features to choose...

Thank you all for your ideas so far!

-- 
Petr Pasky Baudis
A lot of people have my books on their bookshelves.
That's the problem, they need to read them. -- Don Knuth
___
computer-go mailing list
computer-go@computer-go.org
http://www.computer-go.org/mailman/listinfo/computer-go/

[computer-go] Neural networks

[Petr Baudis]

  Hi!

  Is there some high-level reason hypothesised about why there are
no successful programs using neural networks in Go?

  I'd also like to ask if someone has a research tip for some
interesting Go sub-problem that could make for a nice beginner neural
networks project.

  Thanks,

-- 
Petr Pasky Baudis
A lot of people have my books on their bookshelves.
That's the problem, they need to read them. -- Don Knuth
___
computer-go mailing list
computer-go@computer-go.org
http://www.computer-go.org/mailman/listinfo/computer-go/

Re: [computer-go] Neural networks

[Petri Pitkanen]

Neural network tend to work well in those cases where evaluation function is
smooth, like backgammon. Even inbackgammon neural networks do give good
results if situation has possibility of sudden equity changes like deep
backgames and deep anchor games. Top backgammon programs 3-ply search on top
neural network to handle these problems.

I do not know wher neural nets would fit well, perhaps finding invasion
spots?

There has been something on chess neural nets so maybe checking what they
have done?

I am pretty sure there was neural net program in later 80ies?

Petri

2009/10/14 Petr Baudis pa...@ucw.cz

  Hi!

  Is there some high-level reason hypothesised about why there are
 no successful programs using neural networks in Go?

  I'd also like to ask if someone has a research tip for some
 interesting Go sub-problem that could make for a nice beginner neural
 networks project.

  Thanks,

 --
Petr Pasky Baudis
 A lot of people have my books on their bookshelves.
 That's the problem, they need to read them. -- Don Knuth
 ___
 computer-go mailing list
 computer-go@computer-go.org
 http://www.computer-go.org/mailman/listinfo/computer-go/




-- 
Petri Pitkänen
e-mail: petri.t.pitka...@gmail.com
___
computer-go mailing list
computer-go@computer-go.org
http://www.computer-go.org/mailman/listinfo/computer-go/

Re: [computer-go] Neural networks

[Erik van der Werf]

In my opinion NeuroGo was quite succesful with neural networks.
Magog's main strength came from neural networks. Steenvreter uses
'neural networks' to set priors in the Monte Carlo Tree.

Erik


On Wed, Oct 14, 2009 at 2:26 PM, Petr Baudis pa...@ucw.cz wrote:
  Hi!

  Is there some high-level reason hypothesised about why there are
 no successful programs using neural networks in Go?

  I'd also like to ask if someone has a research tip for some
 interesting Go sub-problem that could make for a nice beginner neural
 networks project.

  Thanks,

 --
                                Petr Pasky Baudis
 A lot of people have my books on their bookshelves.
 That's the problem, they need to read them. -- Don Knuth
 ___
 computer-go mailing list
 computer-go@computer-go.org
 http://www.computer-go.org/mailman/listinfo/computer-go/

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

Re: [computer-go] Neural networks

[Rémi Coulom]

Petr Baudis wrote:

  Hi!

  Is there some high-level reason hypothesised about why there are
no successful programs using neural networks in Go?

  I'd also like to ask if someone has a research tip for some
interesting Go sub-problem that could make for a nice beginner neural
networks project.

  Thanks,



At the time when it was fashionable, I would have sold my pattern-Elo 
stuff as a neural network, because, in neural-network jargon, it is in 
fact a one-layer network with a softmax output. Since the development of 
support-vector machines, neural networks have been considered completely 
obsolete in the machine-learning community. From a marketing point of 
view, it is not a good idea to do research on neural networks nowadays. 
You must give your system another name.


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

Re: [computer-go] Neural networks

[Heikki Levanto]

On Wed, Oct 14, 2009 at 03:34:59PM +0300, Petri Pitkanen wrote:
 Neural network tend to work well in those cases where evaluation function is
 smooth, like backgammon. Even inbackgammon neural networks do give good
 results if situation has possibility of sudden equity changes like deep
 backgames and deep anchor games. Top backgammon programs 3-ply search on top
 neural network to handle these problems.
 
 I do not know wher neural nets would fit well, perhaps finding invasion
 spots?

I have been speculating about a NN evaluation function for go, feeding it a
lot of preprocessed information about the position, like number of strings
with 1,2,3,4, or more liberties, number of stones in same, number of separate
groups, number of obviously dead stones, strings, and groups, number of
points clearly controlled by each player, etc, etc. This should be possible
to train from existing games where we know the result (in the beginning it is
50-50, in the end one or the other has won 100-0. Assume some simple function
in between). 

I have never had the time nor the patience to pursue this any further - I
have more interesting ideas, and far too little time... If anyone wants to
play with it, I'd love to hear of any results.

  - Heikki

-- 
Heikki Levanto   In Murphy We Turst heikki (at) lsd (dot) dk

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

Re: [computer-go] Neural networks

[George Dahl]

Neural networks are not considered obsolete by the machine learning
community; in fact there is much active research on neural networks
and the term is understood to be quite general.  SVMs are linear
classifiers for hand-engineered features.  When a single layer of
template-matchers isn't enough, neural networks can be quite effective
for extracting features that could be a kernel for an SVM or
whatnot.  If anything, neural networks research gets marketed as
research on probabilistic graphical models more often than it gets
marketed as kernel machines research.

- George

On Wed, Oct 14, 2009 at 9:06 AM, Rémi Coulom remi.cou...@univ-lille3.fr wrote:
 Petr Baudis wrote:

  Hi!

  Is there some high-level reason hypothesised about why there are
 no successful programs using neural networks in Go?

  I'd also like to ask if someone has a research tip for some
 interesting Go sub-problem that could make for a nice beginner neural
 networks project.

  Thanks,


 At the time when it was fashionable, I would have sold my pattern-Elo stuff
 as a neural network, because, in neural-network jargon, it is in fact a
 one-layer network with a softmax output. Since the development of
 support-vector machines, neural networks have been considered completely
 obsolete in the machine-learning community. From a marketing point of view,
 it is not a good idea to do research on neural networks nowadays. You must
 give your system another name.

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

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

RE: [computer-go] Neural networks - numenta

[David Fotland]

Remi,  what do you think of Numenta http://www.numenta.com/, a startup that
is using feedforward/feedback networks to model learning and pattern
recognition in the neocortex.  Does this approach make sense or is it just
startup hype?

http://www.numenta.com/for-developers/education/biological-background-htm.ph
p


David

 -Original Message-
 From: computer-go-boun...@computer-go.org [mailto:computer-go-
 boun...@computer-go.org] On Behalf Of Rémi Coulom
 Sent: Wednesday, October 14, 2009 6:07 AM
 To: computer-go
 Subject: Re: [computer-go] Neural networks
 
 Petr Baudis wrote:
Hi!
 
Is there some high-level reason hypothesised about why there are
  no successful programs using neural networks in Go?
 
I'd also like to ask if someone has a research tip for some
  interesting Go sub-problem that could make for a nice beginner neural
  networks project.
 
Thanks,
 
 
 At the time when it was fashionable, I would have sold my pattern-Elo
 stuff as a neural network, because, in neural-network jargon, it is in
 fact a one-layer network with a softmax output. Since the development
 of
 support-vector machines, neural networks have been considered
 completely
 obsolete in the machine-learning community. From a marketing point of
 view, it is not a good idea to do research on neural networks nowadays.
 You must give your system another name.
 
 Rémi
 ___
 computer-go mailing list
 computer-go@computer-go.org
 http://www.computer-go.org/mailman/listinfo/computer-go/

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

Re: [computer-go] Neural networks - numenta

[David Doshay]

I'm not Remi, but I know a bit about Numenta. I gave a lightning  
talk at their
workshop about a year and a half ago. A few people at Numenta are  
interested
in using their software for Go, and I was working with one of them  
before my

heart problems stopped that work.

I do not think that they are all start-up hype, but their product is  
still short a
few features that would be needed to play Go. Lately they have mostly  
been
working on image recognition, and their proposed prediction module  
won't
get up to release standards until they are even further along with the  
vision

modules.

The existing image modules could possibly be used for evaluation  
functions,
but that seemed hard to me. I am willing to wait for the prediction  
module

that would more directly lead to move generation.

Cheers,
David



On 14, Oct 2009, at 9:03 AM, David Fotland wrote:

Remi,  what do you think of Numenta http://www.numenta.com/, a  
startup that

is using feedforward/feedback networks to model learning and pattern
recognition in the neocortex.  Does this approach make sense or is  
it just

startup hype?

http://www.numenta.com/for-developers/education/biological-background-htm.ph
p


David


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

Re: [computer-go] Neural networks - numenta

[Rémi Coulom]

David Fotland wrote:

Remi,  what do you think of Numenta http://www.numenta.com/, a startup that
is using feedforward/feedback networks to model learning and pattern
recognition in the neocortex.  Does this approach make sense or is it just
startup hype?

http://www.numenta.com/for-developers/education/biological-background-htm.ph
p


David


Hi,

I don't think I am in a better position than anybody else to reply to 
this question. I am sorry I do not know Numenta, and I do not have time 
to study what it does in details.


Some years ago, I was very interested in studying the brain. My PhD was 
in Cognitive Science, not Computer Science. But I am much less 
interested in that stuff now.


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

Re: [computer-go] Neural networks - numenta

[Mark Boon]

Also not Remi, but...

Numenta is a startup funded by Palm founder Jeff Hawkins. He started
it following up on his book 'On Intelligence', which I think is a very
interesting read. I'd suggest it as a reading to anyone considering
applying some form of Neural simulation to Go or any other problem.

At some point I did try to figure out what Numenta was trying to do.
To some extent it sounds interesting but with limited use. IMO it
strayed quite a bit from the original ideas from Hawkins' book. At
least that is my impression. And what I don't like is that the
solution doesn't translate. So for a pattern-matcher for Go for
example, it has to learn the patterns at all possible board-locations
in all orientations. The abstraction part where you see a pattern on
one part of the board and then being able to recognise it on another
part or in another rotation of the board is missing.

But that doesn't necessarily mean there aren't any interesting ideas
in there. I just think it's not nearly half there to be generally
useful.

Mark


On Wed, Oct 14, 2009 at 6:03 AM, David Fotland fotl...@smart-games.com wrote:
 Remi,  what do you think of Numenta http://www.numenta.com/, a startup that
 is using feedforward/feedback networks to model learning and pattern
 recognition in the neocortex.  Does this approach make sense or is it just
 startup hype?

 http://www.numenta.com/for-developers/education/biological-background-htm.ph
 p


 David

 -Original Message-
 From: computer-go-boun...@computer-go.org [mailto:computer-go-
 boun...@computer-go.org] On Behalf Of Rémi Coulom
 Sent: Wednesday, October 14, 2009 6:07 AM
 To: computer-go
 Subject: Re: [computer-go] Neural networks

 Petr Baudis wrote:
    Hi!
 
    Is there some high-level reason hypothesised about why there are
  no successful programs using neural networks in Go?
 
    I'd also like to ask if someone has a research tip for some
  interesting Go sub-problem that could make for a nice beginner neural
  networks project.
 
    Thanks,
 

 At the time when it was fashionable, I would have sold my pattern-Elo
 stuff as a neural network, because, in neural-network jargon, it is in
 fact a one-layer network with a softmax output. Since the development
 of
 support-vector machines, neural networks have been considered
 completely
 obsolete in the machine-learning community. From a marketing point of
 view, it is not a good idea to do research on neural networks nowadays.
 You must give your system another name.

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

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

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

Re: [computer-go] Neural networks

[dhillismail]

[ Digression: Some ANN architectures are more biologically plausible than 
others. Neuromorphic Engineering is a good search term to see what's going on 
along those lines. (But for a beginner project, the standard multi-layer 
perceptron with backpropogation would still be the natural choice.) Neural nets 
have been called the second best solution to every problem. That's both 
criticism and praise. When the objective is to quickly solve the engineering 
problem at hand and then move on with your life, it can be a handy skill to 
have. ]

 

For an interesting but realistic beginner project, I like this suggestion from 
Magnus.

 
 From: Magnus Persson magnus.pers...@phmp.se



 I guess neural networks is fine for learning pattern priorities for example.

 There are probably just simpler and faster 

 methods for doing that. 
 


 Anyway a good project would be learning 3x3 patterns for MC heavy playouts 
 with a large number of extra features 

 such as exact liberty counts, distance from edge and corner etc. I guess the 
 best use of a neural network would be

 to create a mapping from a huge feature space to a move ordering, where more 
 straightforward pattern matching 

 would fail because the number of patterns are too high. 



One could use the simpler and faster methods to compare against one's own 
results, before moving up to larger patterns, and patterns combined with 
features.



- Dave Hillis

 
Quoting Petr Baudis pa...@ucw.cz: 
 
 Hi! 
 
 Is there some high-level reason hypothesised about why there are 
 no successful programs using neural networks in Go? 
 
 I'd also like to ask if someone has a research tip for some 
 interesting Go sub-problem that could make for a nice beginner neural 
 networks project. 
 
 Thanks, 
 
 -- 
 Petr Pasky Baudis 
 A lot of people have my books on their bookshelves. 
 That's the problem, they need to read them. -- Don Knuth 
 ___ 
 computer-go mailing list 
 computer-go@computer-go.org 
 http://www.computer-go.org/mailman/listinfo/computer-go/ 
 
 
 
--Magnus Persson 
Berlin, Germany 
___ 
computer-go mailing list 
computer-go@computer-go.org 
http://www.computer-go.org/mailman/listinfo/computer-go/ 

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

[computer-go] Neural Networks

[Joshua Shriver]

Anyone recommend a good book on programming Neural Networks in C or C++?

Been digging around the net for while and haven't come up with
anything other than an encyclopedia-like definition/writeup. No
examples or tutorials.

Thanks!
-Josh
___
computer-go mailing list
computer-go@computer-go.org
http://www.computer-go.org/mailman/listinfo/computer-go/

Re: [computer-go] Neural Networks

[George Dahl]

FANN (http://leenissen.dk/fann/) is a great neural network library
written in C.  I don't know much about books on *programming* neural
networks specifically, but I know of many great books on neural
networks.  I am a big fan of Bishop's Neural Networks for Pattern
Recognition even if you aren't necessarily going to use them just for
pattern recognition.
- George

On 7/20/07, Joshua Shriver [EMAIL PROTECTED] wrote:

Anyone recommend a good book on programming Neural Networks in C or C++?

Been digging around the net for while and haven't come up with
anything other than an encyclopedia-like definition/writeup. No
examples or tutorials.

Thanks!
-Josh
___
computer-go mailing list
computer-go@computer-go.org
http://www.computer-go.org/mailman/listinfo/computer-go/


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

Re: [computer-go] Neural Networks

[Richard Brown]

On 7/20/07, Joshua Shriver [EMAIL PROTECTED] wrote:

Anyone recommend a good book on programming Neural Networks in C or C++?

Been digging around the net for while and haven't come up with
anything other than an encyclopedia-like definition/writeup. No
examples or tutorials.


There are some C programs at
http://www.neural-networks-at-your-fingertips.com/
that you may find useful.  [Reading this code to see what it does
helped me to learn!]

Software that can be downloaded from
http://www.dontveter.com/nnsoft/nnsoft.html
is similarly useful, as is Don  Tveter's book The Pattern Recognition Basis of
Artificial Intelligence, to which that software is a companion.
[What I like about
Tveter's book, and software, is that it seems to be written more from
an engineering
perspective than a theoretical one.  As you have found, there is no shortage of
theory, on the web, but somewhat of a dearth of practical information.]

There is also a repository of free AI software (including some for
neural networks) at
http://www.cs.cmu.edu/Groups/AI/html/rep_info/intro.html .

Also, of course, you might wish to browse the archives of the Usenet newsgroup,
available at  http://groups.google.com/group/comp.ai.neural-nets  or from your
friendly neighborhood NNTP server.  The FAQ list, in particular,  for that group
contains a lot of good links.

Hope this helps.

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

Re: [computer-go] Neural Networks

[forrestc]

George Dahl said:
 FANN (http://leenissen.dk/fann/) is a great neural network library
 written in C.  I don't know much about books on *programming* neural
 networks specifically, but I know of many great books on neural
 networks.  I am a big fan of Bishop's Neural Networks for Pattern
 Recognition even if you aren't necessarily going to use them just for
 pattern recognition.

I also see a certain amount of promise in neural networks done wrong.

Current notion [I'm really trying to program this, arrrgh!]

A breeding population of cells, fibers, rules. Each cell has one or more
fibers  one or more rules. Each rule has a threshold and a masked
pattern; if the cell's charge is over the threshold and its pattern
matches the rule's pattern, the cell fires  its pattern goes out via all
fibers. [That is: The pattern  a list of the cell's fibers are added to
the end of a list of fibers to be processed.]

Each fiber has a bit-logical operation as well as a destination cell. When
a fiber is processed, its pattern and operation are applied to the
destination cell's pattern. Then the fiber adds its 'signal strength' to
the cell's charge; and this is checked against the thresholds of the
cell's various rules, and so on...

The initial patterns are, of course, bitmaps of the board. When a nonzero
signal finally arrives at cell #5, one of the 1-bits is randomly selected
as the move. (If that turns out illegal, the actual move is pass.)

Since these cells are randomly connected (depending on the player's
'genes' plus any modifications in the course of the game)... that fibers
to process list could easily blow up; what I think I'll do to damp out
the overload is to add the square of (How many times has this cell fired?)
to the thresholds.

If I get this running, there will be a whole mess of information being
processed about each position... some of it maybe even relevant. Anyway,
putting together a stew of good borrowed ideas... in what I hope will be a
functional way... it might be fun to watch.

Forrest Curo
San Diego


-
This email was sent using AIS WebMail.
http://www.americanis.net/


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

Re: [computer-go] Neural Networks

[Chris Fant]

The initial patterns are, of course, bitmaps of the board. When a nonzero
signal finally arrives at cell #5, one of the 1-bits is randomly selected
as the move. (If that turns out illegal, the actual move is pass.)


Without understanding everything about what you're doing (not even
close), I'm wondering why you wouldn't pick different 1-bit when the
first one you tried was illegal.  And of course pass when they are all
illegal.  Probably you would also have some other mechanism for
generating passes even in the case that they are not all illegal.
___
computer-go mailing list
computer-go@computer-go.org
http://www.computer-go.org/mailman/listinfo/computer-go/