Is the human thought process predictable?

probability

Depends on what you mean by accurately predicted I suppose.

"Probability" is not devoid of prediction.  It is, in fact, a prediction of ALL likely outcomes.

* ALL possible outcomes.  Surely this makes it even more difficult to achieve, not easier than predicting just one outcome?

eh ?

Very deep Betfairy. how come when given the same info,
we can come up with different views ?

But it is impossible to achieve isn't it? No man can accurately predict the result of any future (non-fixed) sporting event can he? I thought we all just guessed, and those that have the best talent for guessing will be those that put the most research into their guess, probably based on history. That's my guess anyway, interesting to hear others views.

The betting patterns of humans is predictable though Betfairy. If you know what the crowd is likely to do, you can use this to your advantage.

Hawking is just looking for excuses as to why he did his brains on Nadal/Melzer imo.

Of course people are predictable. If you go out on a saturday night and come home late all pissed you know exactly what the wife is going to do and say!  Most people have a very narrow band a of awareness with regard to what they are going to do in a situation and likewise sports people.  So once you have seen what they have done before you can predict what they are likely to do again.  This goes for horses as well  and working on that basis profits can be made  betting on that likely outcome

if you could predict one thing, just one event, with 100% probability, then there would be no such thing as Free Will, no such thing as Betting, and, most likely, no such thing as the Universe.

imo

Good try at upping the intellectual level of the forum Betfairy.
Lost me though, I'm afraid,  in the process.

Btw Betfairy I can never predict your thought processes. So QED.

Where does Hawking state this belief in freewill ? , always thought he was a determinist .

I thought he was a dermatologist.

Same thing.

It seems to me that he is not saying that the human thought process is not predictable, but that it is too complex to predict with certainty.

Obviously if I step out into the street and punch a random guy in the face, I can play through a set of scenarios of what is likely to happen. If I first collected background information on the individual I would be able to assign a probability to the events a)run away b)fight back c) start shouting d)make threats etc.

If we could predict things with 100% accuracy, there would be no betting as everything would fall under yes it will happen or no it won't.

Also you might kill him with your first punch.

"to replicate it's workings would require a replica of the universe (or words to that effect)."


IMO everything is constantly changing and the brain is only a physical "calculator" mechanism in that sense for what the non-physical mind plucks from the Universal thought consciousness. So an individuals thoughts and choices would constantly change and adapt through their experiences, making it impossible to "read" them with 100% accuracy.

At the same time that doesnt mean the collective consciousness of the masses engaged in one event, so to speak, wouldnt repeat patterns of learned behaviour, especially with regards to a betting exchange and the fear and greed traits seen everyday for example.

I have always felt that intuition plays an important role in trading the exchanges. Some people stick rigidly to prices and probability to find a margin, but in my view, intuitevly reading the "herd" is also a powerful tool to work with.

The human thought process perceives ambiguous data in erroneous and predictable ways.  The shrewd layer or bettor can exploit this if aware of these biases.

Some people stick rigidly to prices and probability to find a margin, but in my view, intuitevly reading the "herd" is also a powerful tool to work with.

The same thing IMO. Reading the herd, usual over reaction, betting against it etc, etc, etc

The smarties all seem to think they're contrarians.
Not really possible though.

I think one has to be contrary if, say betting on football,
as overall, it's a fact that around 90% of punters lose and
perhaps 90% of them, go with the popular vote, with the rest
being contrary for no other reason than that. After all there
are some matches when it is impossible to back one side
against another, other that to being contrary. Does that make
sense ? Maybe not.

In theory you have to take into account all factors of the universe to calculate the move of the Moon around the Earth or that of the Earth around the Sun.

But Newton did it with a simplified model with a high degree of accuracy.

The same will sooner or later hold true for the human brain, given that the number of neurons, interconnections of neurons and relevant external factors are finite and within human scales (although we still lack the computational power)

Don't hold your breath though.

Nobody really expected computers to take on human chess players so soon and with such a superiority with machines at the reach of our fingertips in any shop ...

Hawking went on a blind date the other night. He came home with a black eye, two cracked ribs and badly grazed knees.

The b1tch had stood him up!

Very good posts saint-pilgrim!

When I was 12 years old one of my teachers said chess computers would never be able to beat humans, as it was humans that provided the input. How could they create something better than themselves? I remember saying that humans had created calculators that could calculate faster and with greater accuracy (or to put it another way: better) than any human, so why wouldn't the same be possible for a chess program. I remember that quite vividly, as it was the first time I really started to question what we were being taught.

One of my favourite quotes is one by George Box: Essentially, all models are wrong, but some are useful.

A model simplifies what is really going on and helps us to understand it. Personally, I believe that an infinite level of complexity exists in nature. We can only understand what is going on by creating artificial boundaries and examining what lies within.

You are really seriously comparing the programming of computers to play chess with the programming of them to replicate the human mind in totality.
Well you read on here first I suppose.

^
That's what I mean by creating artificial boundaries. Parts of what we as humans can do can be replicated by a computer. Calculation, interpreting written text etc.

The more things you add to this list, the closer you get to replicating the human mind in totality. I believe you can eventually get close to it, but you can't actually replicate it due to the 'infinite level of complexity'.

The computer chess solution is a red herring.  The software doesn't replicate the human mind at all - instead it uses a clever mix of memory "book knowledge" and brute force.  The human thought process is much more subtle.

The end results are the same in this case, but that's about all.

A mobile phone won a category 6 tournament with a performance rating 2898. Chess engine Hiarcs 13 running inside Pocket Fritz 4 on the mobile phone HTC Touch HD won the Copa Mercosur tournament in Buenos Aires, Argentina with 9 wins and 1 draw on August 4–14, 2009



The thought processes of Wayne Rooney dont add up to anything meaningful when working out odds on man u matches. Cattermole is always likely to get sent off though and should be factored in to Sunderland odds

Logically there is no difference between the human brain and the computer it has devised. The only difference, and it is a very pure one, is that the human brain is aware of the decisions it makes and the consequences of those decisions. The human has a moral centre. Why evolution has played such a trick is open to debate, but it has. Computers will never gain the capacity for moral understanding.

Never is too strong a word to use in predicting anything.
I prefer probably never.

It's time for the "Worker of the Week Award". I can't believe we've overlooked this week's winner for so very, very long. We simply could not function without his tireless efforts. So, a round of applause for...this inanimate carbon rod!

[Everyone cheers]

.
http://picture.funnycorner.net/funny-pictures/5178/Inanimate-carbon-rod.jpg

The Betfairy Joined: 29 Apr 06
Replies: 7584 15 Oct 10 20:31 
The computer chess solution is a red herring.  The software doesn't replicate the human mind at all - instead it uses a clever mix of memory "book knowledge" and brute force.  The human thought process is much more subtle.

The end results are the same in this case, but that's about all.



No, because the computer can also predict with a high degree of accuracy most of the candidate moves for a high level player, so it shows that within this admittedly limited field the human thought process is highly predictable.

A computer won't undertand exactly how a human player selects candidate moves, which I think is what you're getting at?
Does it matter though?

Why does Kenilworth now post in italics, is he special?

The Betfairy is correct about chess computers.  If they didn't have a massive set of moves and positions pre-programmed, they wouldn't be anywhere near as good even with brute force.  In fact, they used to play in a different style lacking the creativity of humans.  I haven't really tracked them much for a number of years, so they may have improved in that respect.

Computers are still not that good at Go, which although much simpler looking than chess, has far more possible positions and combinations.  Brute force is therefore still a long, long way off mastering Go.

On a more general point, Artificial Intelligence has not made the progress it should have over the years.  Even the Turing Test is not close to being met without bluffing.

"--- within this highlty limited field-----"
I agree at least with that Investor.
I think I'll still go with Hawking's views in this debate.
I think he's probably a bit brighter than me on such matters ( on all matters most likely ).

McChicken, I'm not the only one posting in italics, in case you
haven't noticed. The buttons sre there, so why not play with them ?

i agree

At least I have learned what the Turing Test is today, so that is something.  Before today, I had never heard of it. I looked it up on Wikipedia, but I found the whole thing a bit boring.

That is an interesting statement .Marksman.. I am Happy. How are you feeling today?

Most people would agree that italicised writing should be reserved for quoting people. Ideally we'd all be using the "quote" bb code for that, but it'll never happen.

Why not try a different font in the mean time...

Zola's Back Heel,
Hello Zola, I'm glad you found my statement interesting.  I slept well last night, so I am happy too.  I hope you slept well last night and that, consequently, you will be happy again today.

The best chess players in the world do also memorise openings (up to the first 30 initial moves in some cases), can recognise positions that have been played before and act to what happened in the past and do mechanically play endings (although we tend less and less to see endings, because the opponent can also recognise a lost position)

Chess Masters do only mentally analyse 2 or 3 moves out of the (average) 25-40 available during mid-game.

All the things described above are programmed into the world best chess programs.

The human brain is a complex collection of neurons, interconnections, electrical and chemical signals and some other things. And sometime from now, we'll also be able to develop artificial brains capable of achieving the same complexity, starting from a very basic set of basic components (just like the brain of a child or even that of an embryo)

Neural networks and its theory are still in their infancy but so was medicine (and to some extent it still is) 100 years ago.

As to whether humans have moral (and centre of moral) I tend to believe that humanity has (very slowly) found out that complying with certain moral patterns is better than complying to none. And we, as children, learn to develop brain patterns than conform (in addition to those pre-programmed in our brain to help us with basic survival) to the moral that has helped our societies (I guess that's west/central-european for most of the posters) evolve better than others.

Refusing to see that we will be able to create better versions of ourselves, is like when many denied that flight could be achieved or that we could go to the moon.

Star Trek is an example: science fiction series that talks about warp fields ... someone in real life already developed equations to achieve warp-like travel. And some others have started refining them to decrease the energy needed and ease the process. Engineers will get at some point involved to solve real life implementation problems and ... Yes ... faster than sound was also impossible ...

Chess Masters can analyse dozens of lines of play, up to 20 and 30 moves deep.  They do this because they can recognise what the important moves are, and what they irrelevant ones are.

Of course, latest chess engines may use similar principles now, but it wasn't that long ago that the only option was brute force, and as we know that can extend into billions of combinations very quickly.

Does a computer "play" chess?
I don't know.

But in the real world, chess is a very simple process.  How much more complex is speech recognition (speech understanding)?

And so can computer programs analyze dozens of lines of play (although they admit they never go deeper than 8 moves, even if they could). Only that know and with the advise of those that know how the game is really played (Masters), they just go deep into the 2-3 variations that matter. But also children learn from the masters and program that into their brains as neuron patterns.

Chess is a simple process and it has a unique solution. Once every single variation has been calculated and stored, white will always win. Just a matter of time.

Speech recognition is different in the sense that it deals with "pattern recognition", which is what our brains are best at (recognize faces, races - dogs, humans, etc -. And this is where neural networks come into play (although today the IVR systems that do customer care use customized algorithms to recognise our "yes", "five" and others) It won't take long for neural network theory, combined with algorithms, where speech recognition will be regular business (for years, different models of telephones have supported voice dialling) and will not fail if the speaker happens to have a foreign accent.

But even in this area, we'll be easily able to surpass ourselves, given the limitations of our auditive systems.

Theory and science aside, I'd like to add that seers, mediums, mentalists and similar species have already proven that human behaviour is predictable and many have made fortunes out of it.

Do mentalists predict?  Or do they influence?

saint-pilgrim,

Whilst I agree that we are only very early on in the age of computing, it has also become clear that some basic hurdles are not being overcome in terms of understanding thought.  Chess is just used as a (simple) example for research.  We can only estimate the number of positions and moves possible and the relatively slow advances of computing power do not even begin to get anywhere near "solving" (fully enumerating) chess.  Connect 4 was only solved a few years ago.  I suspect that quantum computing (or some other technology yet to be conceived) will be required to make any breakthroughs.

Incidentally, I haven't seen any evidence of white always winning in an optimal game of chess.  Have you got any links?  Empirically, I would have thought that a draw would be the outcome, as the majority of games between top players are draws.

They don't predict in the broad sense of the word. They predict human behaviour ... because even with a complex neural network and countless (in the non-simplified model) external influences ... we, humans, tend to be very simple ...

Give people the chance to chase their losses ... and they will

If we're so simple, what am I thinking in 5 minutes?

I am no mentalist and if I were I wouldn't be writing here but enjoying myself in the Bahamas after a successful 10 year run in the business with lots of people thinking how it was possible for me to predict what they were thinking in the next 5 minutes.

I would say that there are some chess openings that are better for black, if played between two computers, and these are rarely played at grandmaster level, because the end result will be the same.  Yet, amongst average players, these could be better for white, because it is more difficult for moderate players to defend than attack.  A good example of this is The Fried Liver Attack. Also, Petrov's Defense is such a solid defence, for black, in the hands of a top player, that 1 e4 e5 2Nf3 has become less common at the top level nowadays.  (This is according to a top female grandmaster on chess dot com.)  Of course, I'm not much good at chess myself and never will be.  I just have some knowledge of it.  I know that Mythical Prince, on the ante post forum, fancies himself as a player, so he is the one to ask about chess.

Petrov is nice if you want look for a draw and avoid any mid-game complexity, whilst some others let you enter more complex and obscure mid-game situations where the white advantage can not be so well used.

But this is only so because human players don't know the exact right combination to follow to always win when playing white. Chess is a game of perfect information and the fact that white move first makes the difference. The problem of chess will be solved and surely some brilliant and talented minds will be able to gain advantage by using the already pre-computed solution (by memorising the first 50 moves for example) to get them into a very advantageous position.

That doesn't mean I will quit playing chess (at my level)

An individual or group response to s set of set stimuli can be predicted to some extent. You only have to read a yet experimental psychology journals to see this. Humans are inexact calculating machines, we constantly look for shortcuts.

Which is why we have learned responses. It would be too exhausting to compute everything from scratch each time. So your brain writes "subroutines" for most repetitive tasks which it engages whenever you have to.

Ask yourself this, which side of your face do you start shaving first? I'd bet its the same side each time (unless you are in a hotel and the lighting is different). Which shoe do you put on first? There are lots of example where your specific behaviour will be predictable to someone who knows what to look for, yet you probably don't even realise it yourself.

Now... *HOW* an individual actually computes their responses/actions and how to mimic how the brain interprets and stores all this innformation is a entire different and vastly more complicated area. 

I'd argue there is no one set algorithm for "tying your shoeslaces". There may be a mathematically optimal or locally least-error appraoch, but if we could dissect exactly how I'd expect everyone to have a slightly different approach.

saint-pilgrim    I don't accept that, with perfect play, white will always win against Petrov's Defense.  This is not proven.  In the queen swap variation, black is 2 moves behind, yet the defence was apparently proved impregnable by Petrosian (black) against Spasky in the world championship.  I have borrowed my brother's book on chess openings and it uses that game as a demonstration that the position is equal for both sides.

In fact I am going there now, and if I am black I will be use Petrov to get at least a draw...

Marksman,

I also have books of openings and know that Petrov gives you a solid defense as I also pointed out. What I said is that once the problem of chess is solved (the number of positions is finite and it will therefore be solved) and given the fact that white always moves first, white will always be able to dictate the result of a match. The result may end up being that black is capable of not losing, but white will always be able to force at least a draw (although I would be surprised if the result isn't that white can always achieve a win)

Of course if humans play (and we have limited computing capacity) ... it's a different story. That's why I also enjoy playing chess.

---

Very nice post Zola!

I am white ; queens gambit accepted

I win.

There are surely regularities in group responses along the lines of what Zola suggests, and the exploitation of these when they are sub-optimal must be central to winning at betting (especially e.g. IR trading).

My sense re Hawking's comment on free will is that if we know every component of a system (e.g. the brain) and every manner in which any component interact with any others, then we will be able to replicate the system. By 'free will' I take Hawking to mean 'we cannot replicate neural systems yet'. But I'm not sure that he's making a point about complexity to the effect that he thinks we'll never be able to simulate / replicate the brain.

imho the stumbling block for the thought processes of a computer to rival that of a human is the fact humans feel emotion. What one person feels about a certain situation is influenced by a multitude of external factors that I think would be very difficult to replicate within any neural network system.

How would all contributors to this thread react to a very serious confrontational situation? How would these thought processes be replicated on a computer? Long way to go yet imo.

There may be "a long way to go" but that does not mean it is not possible. The brain in a highly non-linear, multi-noded self-referential feedback system. Even if we knew exactly how everything linked together and interacted - slight variations in input could lead to vastly different outcomes in a handle-turning computation.

However... I expect the brain has a lot of grey-area (no pun intended) in its OS which filters out a lot of computational jitter which modern computers as yet cannot. If anyone has had experience simulating multi-dimentional highly non-linear systems on massively parallel machines will know what a pain the smoothing fudge factors you have to use to stabalise them [:p].

Free will could just be a initialisation problem.

Emotion is again just a learned response, which we have found useful to develop for social situations. I see no reason why a significantly complex computer could not replicate emotion. In fact I see us just as nothing more than biological computers anyway!

I agree it is certainly possible, but I do believe we are still very far away from achieving a breakthrough of this level.

I do fully agree with Zola. Emotions are just developed (also some ROM programming to let us face unknown situations and let the human race remain on the move) patterns to let us adapt to situations.

And this adaptation patterns can be programmed.

The day we make computers complex enough to be better than ourselves (call it ... the next step in evolution), we'll have to add "emotional responses" (to some extent) to let the better-humans be able to react and adapt to unexpected situations and be able to develop a sensible behaviour, in case the situations repeat.

But who is adding the emotional responses? Will they not be subject to their own emotional responses about which emotional responses are better than others?

The field of psychology is still developing and so, going back to my original point, we are miles off a situation like this.

One of the issues is going to be that humans like to define themselves to be special, so the definition of something that is capable of independent thought is going to be very rigorous indeed (until the machines define it to be something that they can do and we can't )

Geez God must be falling off his chair with laughter if he's monitoring this thread.

I bet you that any useful replication of the brain by compuetrs happens on Wall St first ( if and when it ever doeas happen ).

The replication in Wall Street would have to be that of a moral-less brain, so the programmers would have less work to do, because this is basically the most efficient way to make money there.

Nice one saint-p.
No morals on Wall St. that's for sure.

A moral-less brain by whose standards? Everyone has a unique perception to morals as everyone has a unique trigger point to emotions, so I really dont see how that can work unless you customise it to a one-size fits all assumption of morals, which kind of defeats the object. It might make money but it isnt a replication as is implied.

How do you decide what is a moral trade on wall street anyway? Likewise, how do you determine which perception of reality is correct on a thread like this? You cant IMO, as it is all perception and what you see and sense becomes your experience from which to work from, which has to be to some extent unique for every individual, or there would really be no point in life.

If living things were created the same, how would any of them experience any form of contrast, from which to grow and learn? So if this is true, surely you can only replicate one individual, which only remains a replica in the now moment, and becomes a distorted one from that point on as the living being changes.

You are right Positive. Possibly the following moral standards apply:

-- In a regular office, the % of people that would execute an action knowing it may (possibly will) the life of a colleague (or that of a 3rd party), no matter how many times they have gone out together, have had dinner together with the wifes and similar actions, is X.

-- In a stock market trading environment (never met Jerome Kerviel, but know a few that operate with large amounts), the % of people is Y.

And in my personal experience Y is orders or magnitude greater than X.

But of course, that implies no judgement of which moral may be the best ...

"it may ruin the life of a colleague ..."

Thats why algorithmic trading in the new goose. Y can be 100%.