AlphaZero is not like other chess programs

[corres]

[quote="Henk"]

Problem with this deep learning is getting training data. If you can extract the training data automatically then that problem is solved.
I remember I was already busy with neural networks in 1993-94. But when I found out I had to collect a great many training examples fun was over.

[/quote]

I feel for you, to get billion single data is a very boring thing.
Team of AlphaZero make lighter and make faster getting data.
They generate data (chess positions) for learning chess and developing chess knowledge of their system.

[Cardoso]

We can say that Alpha Zero is a nominalist chess engine?

AlphaZero is not a chess engine.
It is a modern computer system dedicated to solve complicated problems like playing go, chess, shogi etc.
Systems like AlphaZero can be used for solving problems of scientific, economic, military, social behavior (even if political issues).....

...economic
...social behaviour
...political
That's a LOT of faith you put on this system, you better lower your expectations a bit, otherwise you might get disapointed.

[corres]

[quote="Lyudmil Tsvetkov"]

Everything is comparable, there is no such thing in the Universe that is not comparable.
Alpha is a 2800 engine, making use of tremendous power.

[/quote]

Really comparable, but there are lot of situation when the comparison are meaningless.
I think you underrate the potential of AlphaZero. If Google would yield enormous mass money with developing chess machine they would do this and Elo of AlphaZero would be much more high.

[corres]

[quote="Cardoso"]

That's a LOT of faith you put on this system, you better lower your expectations a bit, otherwise you might get disapointed.

[/quote]

These are not my faith, these are the expectations of Google.

[shrapnel]

This "zero" approach really surprised me, starting from almost nothing except the game's rules. I wasn't expecting this advancement so soon, even though they used pretty powerful hardware.

One of the things that amazes me is that Alpha Zero isn't blind to deep tactics (taking into consideration it only runs at 80K positions per second).
Can someone try to explaine this?
I mean I can understand it plays well positionally, but also being tactical so strong is surprising to me.

The deep neural network connects the pieces on different squares to each other. They use 3x3 convolutions. This means that the next 8x8 layer's cells are connected to a 3x3 region (called "receptive field") in the previous region, and to a 5x5 region in the layer before etc. After only 4 layers, each cell is connected to every other cell in the original input layer. For AlphaGoZero they used no less than 80 layers. Then they also have many "feature maps" in parallel, so that they can learn different concepts related to piece-square combinations. Finally, they use the last 8 positions as input as well, so they also have a sense of ongoing maneuvers. All this is then being trained on the game result and the best move from the MC tree search.

Although the amount of resources required to train the millions of weights related to these neural networks is enormous, conceptually it is not surprising that pawn structure, king safety, mobility and even deep tactics can be detected from the last 8 positions.

That's the scientific explanation for you, Cardosa.

[shrapnel]

This "zero" approach really surprised me, starting from almost nothing except the game's rules. I wasn't expecting this advancement so soon, even though they used pretty powerful hardware.

One of the things that amazes me is that Alpha Zero isn't blind to deep tactics (taking into consideration it only runs at 80K positions per second).
Can someone try to explaine this?
I mean I can understand it plays well positionally, but also being tactical so strong is surprising to me.

Quite simple.
Scientific explanation others can give you. I can just tell you what it means.
Positional Play, Tactical Play.... these are just Terms used by humans and obsolete chess engines like Stockfish, Houdini and Komodo.
The AlphaZero Program, after learning Chess in just a few Hours, has one and only one Goal, to Mate the Enemy King, NOTHING else !
What you see as Positional or Tactical Play, means NOTHING to this Monster Program. It wants to win and only win !
Positional or Tactical Play, is only a means to an end, for this Program.
It doesn't care what you call the type of Moves it makes.
It just makes the Moves it needs to make, to achieve its goal of winning.
Simple but deadly.

I'd just like to add a little to my Post to make things clearer.
I ran some of the sample games between AlphaZero and Stockfish. I used Komodo and Houdini and tried to see if either could find some of the Tactical Brilliancies played by AlphaZero, already commented on by others in this Forum.
While Komodo was completely unable to emulate those Tactical Brlliancies, Houdini 6.03, while run in Tactical Mode did manage to discover some of those brilliancies, albeit with a few Tweaks.
Initially, I was thrilled, but then I tried out some of the Games where people said that AlphaZero had played brilliant Positional moves.
Unfortunately, Houdini 6.03 in Tactical Mode was unable to replicate those moves.
So, in essence, AlphaZero seems to instinctively know when to make Tactical moves and when to search deeper for a good positional move.
Awesome, for a Program which just self-trained for a few hours, don't you think ?

[shrapnel]

We can say that Alpha Zero is a nominalist chess engine?

If it runs on a computer system and is a software, I guess you could call it a chess engine.
But that's the only similarity with engines like Houdini and Stockfish, since it uses revolutionary methods to find the correct move, which Houdini and Stockfish can only dream of !

[mclane]

Stockfish and Komodo and Houdini play machine chess.
AZ plays a human like chess.

It sacs pieces for attack or development or space.

And the sacs do NOT lead to a material compensation soon, otherwise
Stockfish and Komodo and Houdini would see the sac and not eat the piece.

It seems AZ plays chess out of the interval stockfish/Komodo/ Houdini searches.


This is very funny to replay.
Stockfish gets smashed down like an idiot.

I guess any human chess player no matter which level he plays can observe that stockfish has no chance at all to win.

The way AZ plays is heavily different then that of the normal chess programs that I feel sorrow for them.
They all play machine chess in their interval of 20-30 plies they search.
But outside this interval, AZ kills them with very easy moves.

A human mind can understand those moves, but for a chess program with a search tree it seems those moves are very difficult to understand.

We see human chess beat machine chess.

With AZ playing like a machine emulated human.

[CheckersGuy]

I hope that AlphaZero will make similar progress as the previous AlphaGo version did.
Maybe they have a version 1 year from now that's considerably stronger than the current one

[corres]

[quote="mclane"]

Stockfish and Komodo and Houdini play machine chess.
AZ plays a human like chess.
It sacs pieces for attack or development or space.
And the sacs do NOT lead to a material compensation soon, otherwise
Stockfish and Komodo and Houdini would see the sac and not eat the piece.
It seems AZ plays chess out of the interval stockfish/Komodo/ Houdini searches.
This is very funny to replay.
Stockfish gets smashed down like an idiot.
I guess any human chess player no matter which level he plays can observe that stockfish has no chance at all to win.
The way AZ plays is heavily different then that of the normal chess programs that I feel sorrow for them.
They all play machine chess in their interval of 20-30 plies they search.
But outside this interval, AZ kills them with very easy moves.
A human mind can understand those moves, but for a chess program with a search tree it seems those moves are very difficult to understand.
We see human chess beat machine chess.
With AZ playing like a machine emulated human.

[/quote]

The team of AlphaZero stated that their machine taught itself from zero knowledge of chess using only the rule of chess as start point.
In your opinion what is the explanation for the phenomenon described by you above? "Thinking" of AlphaZero is based on vectors and probability.
This is very different from thought of human.