[Computer-go] 9x9 is last frontier?

Brian Sheppard sheppardco at aol.com
Wed Mar 7 09:45:27 PST 2018


The algorithms presented there continue to the end of the game tree, no? (I am looking specifically at Algorithm 12, which seems to be the paper’s recommended algorithm.)

 

I am trying to get my head around what happens in practical settings, were we decide to limit search effort. E.g., if the search is limited by depth, then I think the UCT bounds that are present as tiebreakers would seldom be used? (Because moves at the root would not have “infinities.”) That is, the approach would simplify to MT-SSS*.

 

It is my impression that iterative-deepening/aspiration-window is the winner among the alpha-beta family of algorithms, so I am curious to learn how algorithm 12 compares. The problem is that iterative-deepening and aspiration-window are hard to analyze. SSS* is more amenable to proofs. A comparison to SSS* is maybe not relevant.

 

Another aspect that I am trying to understand is whether the associative memory is really an improvement over a transposition table. Specifically: Algorithm 12 says “traverse the move that has the highest upper bound”, where a chess program would have stored a move in the xref table.

 

All in all, if I wanted to use alpha-beta in a Go program, then I would first train a value/priority network as described in the AGZ/AZ papers. Then use it in an alpha-beta program and start researching game-specific search techniques. Singular extensions, late-move-reductions, null-move, …

 

From: Computer-go [mailto:computer-go-bounces at computer-go.org] On Behalf Of Dan
Sent: Tuesday, March 6, 2018 5:55 PM
To: computer-go at computer-go.org
Subject: Re: [Computer-go] 9x9 is last frontier?

 

Alpha-beta rollouts is like MCTS without playouts (as in AlphaZero), and something that can also do alpha-beta pruning.

 

With standard MCTS, the tree converges to a minmax tree not an alpha-beta tree, so as you know there is a huge branching factor difference there.

 

For MCTS to become competitive with alpha-beta, one has to use either alpha-beta rollouts OR as in AlphaZero's case

pick accurately, say the best 3 moves with its PUCT bias, and then search those with MCTS that would converge to minimax (not alpha-beta).

 

I went for the latter (recasting alphabeta in rollouts form which can be combined with standard MCTS as well as standard recursive alpha-beta implementation)

 

The paper is here https://www.microsoft.com/en-us/research/wp-content/uploads/2014/11/huang_rollout.pdf

 

If it was upto me, I would have used standard alpha-beta search with policy/value networks to improve move sorting and evaluation resp.

 

It seems AlphaZero's MCTS approach hinged on the policy network accurately picking the best n moves (while still using minmax) AND then

ofcourse not making blunders. I think what "shallow traps" are some bad looking moves turning out to be brilliant tactically later. How the policy

nework able to identify those moves ? A full width like alpha-beta does not miss those.

 

Daniel

 

 

 

On Tue, Mar 6, 2018 at 1:23 PM, Álvaro Begué <alvaro.begue at gmail.com <mailto:alvaro.begue at gmail.com> > wrote:

Sorry, I haven't been paying enough attention lately to know what "alpha-beta rollouts" means precisely. Can you either describe them or give me a reference?

Thanks,
Álvaro.



 

On Tue, Mar 6, 2018 at 1:49 PM, Dan <dshawul at gmail.com <mailto:dshawul at gmail.com> > wrote:

I did a quick test with my MCTS chess engine wth two different implementations.

A standard MCTS with averaging, and MCTS with alpha-beta rollouts. The result is like a 600 elo difference

 

Finished game 44 (scorpio-pmcts vs scorpio-mcts): 1/2-1/2 {Draw by 3-fold repetition}

Score of scorpio-mcts vs scorpio-pmcts: 41 - 1 - 2  [0.955] 44

Elo difference: 528.89 +/- nan

 

scorpio-mcts uses alpha-beta rollouts

scorpio-pmcts is "pure" mcts with averaging and UCB formula.

 

Daniel

 

On Tue, Mar 6, 2018 at 11:46 AM, Dan <dshawul at gmail.com <mailto:dshawul at gmail.com> > wrote:

I am pretty sure it is an MCTS problem and I suspect not something that could be easily solved with a policy network (could be wrong hree). My opinon is that DCNN is not

a miracle worker (as somebody already mentioned here) and it is going to fail  resolving tactics.  I would be more than happy with it if it has same power as a qsearch to be honest.

 

Search traps are the major problem with games like Chess, and what makes transitioning the success of DCNN from Go to Chess non trivial.

The following paper discusses shallow traps that are prevalent in chess. ( https://www.aaai.org/ocs/index.php/ICAPS/ICAPS10/paper/download/1458/1571 )

They mention traps make MCTS very inefficient.  Even if the MCTS is given 50x more time is needed by an exhaustive minimax tree, it could fail to find a level-5 or level-7 trap.

It will spend, f.i, 95% of its time searching an asymetric tree of depth > 7 when a shallow trap of depth-7 exists, thus, missing to find the level-7 trap.

This is very hard to solve even if you have unlimited power.

 

The plain MCTS as used by AlphaZero is the most ill-suited MCTS version in my opinion and i have hard a hard time seeing how it can be competitive with Stockfish tactically.

 

My MCTS chess engine with  AlphaZero like MCTS was averaging was missing a lot of tactics. I don't use policy or eval networks but qsearch() for eval, and the policy is basically

choosing which ever moves leads to a higher eval.

 

a) My first improvement to the MCTS is to use minimax backups instead of averaging. This was an improvmenet but not something that would solve the traps

 

b) My second improvment is to use alphabeta rollouts. This is a rollouts version that can do nullmove and LMR etc... This is a huge improvment and none of the MCTS

versons can match it. More on alpha-beta rollouts here ( https://www.microsoft.com/en-us/research/wp-content/uploads/2014/11/huang_rollout.pdf )

 

So AlphaZero used none of the above improvements and yet it seems to be tactically strong. Leela-Zero suffered from tactical falls left and right too as I expected.

 

So the only explanation left is the policy network able to avoid traps which I find hard to believe it can identify more than a qsearch level tactics.

 

All I am saying is that my experience (as well as many others) with MCTS for tactical dominated games is bad, and there must be some breakthrough in that regard in AlphaZero

for it to be able to compete with Stockfish on a tactical level.

 

I am curious how Remi's attempt at Shogi using AlphaZero's method will turnout.

 

regards,

Daniel

 

 

 

 

 

 

 

 

On Tue, Mar 6, 2018 at 9:41 AM, Brian Sheppard via Computer-go <computer-go at computer-go.org <mailto:computer-go at computer-go.org> > wrote:

Training on Stockfish games is guaranteed to produce a blunder-fest, because there are no blunders in the training set and therefore the policy network never learns how to refute blunders.

 

This is not a flaw in MCTS, but rather in the policy network. MCTS will eventually search every move infinitely often, producing asymptotically optimal play. But if the policy network does not provide the guidance necessary to rapidly refute the blunders that occur in the search, then convergence of MCTS to optimal play will be very slow.

 

It is necessary for the network to train on self-play games using MCTS. For instance, the AGZ approach samples next states during training games by sampling from the distribution of visits in the search. Specifically: not by choosing the most-visited play!

 

You see how this policy trains both search and evaluation to be internally consistent? The policy head is trained to refute the bad moves that will come up in search, and the value head is trained to the value observed by the full tree. 

 

From: Computer-go [mailto:computer-go-bounces at computer-go.org <mailto:computer-go-bounces at computer-go.org> ] On Behalf Of Dan
Sent: Monday, March 5, 2018 4:55 AM
To: computer-go at computer-go.org <mailto:computer-go at computer-go.org> 
Subject: Re: [Computer-go] 9x9 is last frontier?

 

Actually prior to this it was trained with hundreds of thousands of stockfish games and didn’t do well on tactics (the games were actually a blunder fest). I believe this is a problem of the MCTS used and not due to for lack of training. 

 

Go is a strategic game so that is different from chess that is full of traps.     

I m not surprised Lela zero did well in go.

 

On Mon, Mar 5, 2018 at 2:16 AM Gian-Carlo Pascutto <gcp at sjeng.org <mailto:gcp at sjeng.org> > wrote:

On 02-03-18 17:07, Dan wrote:
> Leela-chess is not performing well enough

I don't understand how one can say that given that they started with the
random network last week only and a few clients. Of course it's bad!
That doesn't say anything about the approach.

Leela Zero has gotten strong but it has been learning for *months* with
~400 people. It also took a while to get to 30 kyu.

--
GCP
_______________________________________________
Computer-go mailing list
Computer-go at computer-go.org <mailto:Computer-go at computer-go.org> 
http://computer-go.org/mailman/listinfo/computer-go


_______________________________________________
Computer-go mailing list
Computer-go at computer-go.org <mailto:Computer-go at computer-go.org> 
http://computer-go.org/mailman/listinfo/computer-go

 

 


_______________________________________________
Computer-go mailing list
Computer-go at computer-go.org <mailto:Computer-go at computer-go.org> 
http://computer-go.org/mailman/listinfo/computer-go

 


_______________________________________________
Computer-go mailing list
Computer-go at computer-go.org <mailto:Computer-go at computer-go.org> 
http://computer-go.org/mailman/listinfo/computer-go

 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://computer-go.org/pipermail/computer-go/attachments/20180307/09707d73/attachment-0001.html>


More information about the Computer-go mailing list