DeepMind's AlphaZero plays chess like a tornado in the junkyard

In situ · Dec 17, 2017

Kenny'sID said:
That is the most ridiculous part of the whole thing...because chess.com says that, it's true?

No, but the quote captures the essence of my point in the research paper written about AlphaZero.

Kenny'sID said:
Give that robot those parts and talk to me afterwards.

Maybe you can have a look at the paper and add your own thoughts afterwards?

In situ · Dec 17, 2017

Tom Farebrother said:
I’m not sure if pattern learning algorithms are strictly analogous to the random generation of mutation

Perhaps not in general but the Monte Carlo Tree Search which AlhpaZero uses is a method to implement reinforced learning used by artificial intelligence.

Kenny'sID · Dec 17, 2017

In situ said:
No, but the quote captures the essence of my point in the research paper written about AlphaZero.

OK, good.

In situ said:
Maybe you can have a look at the paper and add your own thoughts?

I've already added my thoughts...prove it. If they are telling the truth, shouldn't be that tough to do.

Tom 1 · Dec 17, 2017

Speedwell said:
The "volition" is the constant production of randomly distributed somatic variation. The "memory" is the storage of the genetic record of reproductively successful variants in the gene pool.

Had to look that up. I can see that the first is a process. Are you saying then that the fact that there is a process is sufficient explanation for why that process takes place? Here for example the analogy is a piece of software, something that is coded to interact with hardware and make it ‘go’. What is it that makes the process you described ‘go’? I’ve come across descriptions of dna as a code, but coupled with a kind of disclaimer that this is just a metaphor, but what does that metaphor, in the scientific worldview, describe? A code makes sense, as a code could be said to drive the hardware of life, but code has to be written. What does the writing of the code in the scientific view?

Tom 1 · Dec 17, 2017

In situ said:
Perhaps not but the Monte Carlo Tree Search which AlhpaZero uses is.

Is it? As described in the link:
The focus of Monte Carlo tree search is on the analysis of the most promising moves, expanding the search tree based on random sampling of the search space.
This is designed and purposeful, a code that is designed for a particular type of purpose, and has a kind of intelligence in the form of analysis algorithms. It isn’t self generated, and when it has completed its purpose, it either does something else it was coded to do, or it stops. I assume you don’t think that random mutation is designed, so what would be the explanation for why it happens - not the process of how it happens, or the fact that it does happen, but why it does, what is the driving force behind it, in your view?

In situ · Dec 17, 2017

FrumiousBandersnatch said:
Indeed, but you haven't mentioned the really impressive part

True, but I did not felt it was relevant to mention it.

In situ · Dec 17, 2017

Tom Farebrother said:
I’m not sure how loose you can get and still have a connection between a pre-designed ability to learn in a specific fashion within clearly defined parameters, and a process that involves randomness leading to improbable outcomes leading to modified randomness (putting it a bit crudely I know) but it would be interesting to have that explained by someone who understands it better than I do.

I am not sure exactly what it is you wonder about, but here is an attempt to an answerL

We know from machine learning that you do not need to know the purpose, i.e the global goal, in order to find a solution. A local, i.e. a clueless, process can home into a global solution without even being aware of doing so. Talking in general terms this is achieved by following a gradient of some sort. Self-organization in biology uses chemical gradients, e.g. ant hives, embryonic development etc, but evolution uses a more subtle gradient which is none physical but nevertheless real since the gradient is defined by the laws of nature, e.g. you need wings to fly, and the interaction between the living and the environment. This gradient exist in an abstract space known as the fitness landscape.

Tom 1 · Dec 17, 2017

In situ said:
I am not sure exactly what it is you wonder about, but here is an attempt to an answerL

We know from machine learning that you do not need to know the purpose, i.e the global goal, in order to find a solution. A local, i.e. a clueless, process can home into a global solution without even being aware of doing so. Talking in general terms this is acheived by following a gradient of some sort. Self organization in biology uses chemical gradients, e.g. ant hives, embryonic development etc, but evolution uses a more subtle gradient which is none physical but nevertheless real since the gradient is defined by the laws of nature, e.g. you need wings to fly, and the interaction between the living and the environment. This gradient exist in an abstract concept of a fitness landscape.

Yes that makes sense in terms of a process. In the analogy, that process is driven by a code that randomly tries actions, analyses the results and learns from that, if that is a sufficient description. What is the code or driving force in evolution? I mean not he process or the fact of it but what drives it.

Speedwell · Dec 17, 2017

Tom Farebrother said:
Had to look that up. I can see that the first is a process. Are you saying then that the fact that there is a process is sufficient explanation for why that process takes place?

You mean what drives it? Thermodynamics, just like any other chemical reaction.

Here for example the analogy is a piece of software, something that is coded to interact with hardware and make it ‘go’. What is it that makes the process you described ‘go’? I’ve come across descriptions of dna as a code, but coupled with a kind of disclaimer that this is just a metaphor, but what does that metaphor, in the scientific worldview, describe? A code makes sense, as a code could be said to drive the hardware of life, but code has to be written. What does the writing of the code in the scientific view?

The continuous operation of the variation/selection feedback loop.

In situ · Dec 17, 2017

Tom Farebrother said:
It’s late here, I’m tired, and it might sound like I’m being facetious but I’m not, I am genuinely interested in this. The analogy, I think, falls down in a few different ways. One of them is the matter of will. So far at least, AI has no will.

A will kind of implies consciousness. Perhaps with 'will' you are referring to 'intentions'? There are machines, i.e. agents, which have intentions. Intention does not require and agent to be consciousness as far as I understand, but you may judge it differently than me and in some cases I've started to doubt myself weather or not we already made conscious robots....

Tom Farebrother said:
It will accomplish (or not) what it has been designed to do, within set parameters, and then stop. Just stop. It won’t decide it wants to do something else, or rehash what it did and do it better just for the sake of it etc.

You be surprised what AI can do these day then...

However I think the claims made are hyped. I read their research paper and I think the conclusion the robots passed the self-awareness test is flawed - for many reasons. Anyway the point is that these robots demonstrates intention.

And if you haven't seen what Boston Dynamics are up to, this might impress on you .

These robots are not just acting autonomous, they are also re-acting to the unpredictable word around them.

Tom 1 · Dec 17, 2017

In situ said:
A will kind of implies consciousness. Perhaps with 'will' you are referring to 'intentions'? There are machines, i.e. agents, which have intentions. Intention does not require and agent to be consciousness as far as I understand, but you may judge it differently than me and in some cases I've started to doubt myself weather or not we already made conscious robots....

You be surprised what AI can do these day then...

However I think the claims made are hyped. I read their research paper and I think the conclusion the robots passed the self-awareness test is flawed - for many reasons. Anyway the point is that these robots demonstrates intention.

And if you haven't seen what Boston Dynamics are up to, this might impress on you .

These robots are not just acting autonomous, they are also re-acting to the unpredictable word around them.

I’ll have to check it out some time. Bit of a part time curiosity for me rather than anything in depth, hence all the questions.

In situ · Dec 17, 2017

Speedwell said:
The "goal" of evolution is that creatures remain fit to survive in changing environments.

The goal, or "the meaning with life", is reproduction, i.e to have babies. If it was survival, as you suggest, then you and I would not be around and talk about it, instead there would be some guys a few billion years old talking about evolution. Therefore fitness is not about survival but a measurement of reproductive success.

In situ · Dec 17, 2017

Tom Farebrother said:
I’ll have to check it out some time. Bit of a part time curiosity for me rather than anything in depth, hence all the questions.

I think one should because there is a revolution in AI just around the corner and most people are not even aware of what is coming in the next decades.

Speedwell · Dec 17, 2017

In situ said:
The goal, or "the meaning with life", is reproduction, i.e to have babies. If it was survival, as you suggest, then you and I would not be around and talk about it, instead there would be some guys a few billion years old talking about evolution. Therefore fitness is not about survival but a measurement of reproductive success.

Is not reproductive success the essence of fitness of an evolving creature? And how is survival not a contributor to it?

In situ · Dec 17, 2017

Astrophile said:
Would this be analogous with God creating the first self-replicating molecules and leaving them to learn how to evolve by random trial and error, or would it be more like God creating a large number of organic molecules and leaving them to develop into life-forms that could evolve, also by trial and error?

It would be analogues to theistic evolution as I see it, yes. But how I cannot tell since I have not thought that through. The analogue is more at the theoretical level of evolutionary theory, i.e. both are using local algorithms depending on randomness when traversing a global search space for a solution .

In situ · Dec 17, 2017

Speedwell said:
Is not reproductive success the essence of fitness of an evolving creature?

Yes.

Speedwell said:
And how is survival not a contributor to it?

Why would survival matter for fitness? It only matter that you can survive long enough to raise offsprings which in turn can bring new offsprings. What happen after that is irrelevant for fitness. Most living being dies after they reproduced. If you want to see a brutal example of this, check out the destiny of the male praying mantis.

Speedwell · Dec 17, 2017

In situ said:
It would be analogues to theistic evolution as I see it, yes. But how I cannot tell since I have not thought that through. The analogue is more at the theoretical level of evolutionary theory, i.e. both are using local algorithms depending on randomness when traversing a global search space for a solution .

It is the randomness of the search which is thought to insulate the solution, to form a barrier between it and divine providence, thus denying God's creativity. It's a metaphysical issue, really, one which the rudimentary formal theology of the creationist Protestant denominations is unequal to.

Speedwell · Dec 17, 2017

In situ said:
Yes.

Why would survival matter for fitness? It only matter that you can survive long enough to raise offsprings which in turn can bring new offsprings.

That's exactly the "survival" I was talking about. I certainly did not mean to imply that the lengthy survival of any individual of a species was a "goal" of evolution.

In situ · Dec 18, 2017

Tom Farebrother said:
In the analogy, that process is driven by a code that randomly tries actions, analyses the results and learns from that, if that is a sufficient description.

Not sure what you mean with 'code', but my initial reaction is "no that is not how really how works", which tells me I probably did not understand you properly.

Tom Farebrother said:
What is the code or driving force in evolution? I mean not he process or the fact of it but what drives it.

On a generic abstract level: the gradient in the fitness landscape.

If you are on a gradient you cannot avoid to move along the gradient. I don't like the picture of the fitness landscape with hills you need to climb. A hill suggest some kind of force are needed to push you up the hill, while rater the opposite is the case; you are falling down into a valley. The gradient works more like an energy minimum; if you put something in the empty air it will fall back to the ground because that is the direction in which the gravitational gradient points.

In situ · Dec 18, 2017

Speedwell said:
It is the randomness of the search which is thought to insulate the solution, to form a barrier between it and divine providence, thus denying God's creativity. It's a metaphysical issue, really, one which the rudimentary formal theology of the creationist Protestant denominations is unequal to.

I kind of lost you there, I am not used to think in terms of such elaborate colorful language. Don't take me wrong, I think it was beutiful written, and I can sense it carries a lot of meaning , but it was written far beyond my own comprehension.

DeepMind's AlphaZero plays chess like a tornado in the junkyard

in vivo veritas

in vivo veritas

Well-Known Member

Optimistic sceptic

Optimistic sceptic

in vivo veritas

in vivo veritas

Optimistic sceptic

Well-Known Member

in vivo veritas

Optimistic sceptic

in vivo veritas

in vivo veritas

Well-Known Member

in vivo veritas

in vivo veritas

Well-Known Member

Well-Known Member

in vivo veritas

in vivo veritas

Similar threads