Computer Software to Predict the Unpredictable

Amigan writes "Professor Jerzy Rozenblit at the University of Arizona was awarded $2.2Million to develop software to predict the unpredictable — specifically relating to volatile political and military situations." From the article: "The software will predict the actions of paramilitary groups, ethnic factions, terrorists and criminal groups, while aiding commanders in devising strategies for stabilizing areas before, during and after conflicts. It also will have many civilian applications in finance, law enforcement, epidemiology and the aftermath of natural disasters, such as hurricane Katrina."

Comments

computer?

[onemorehour]

Sure, no problem. The software should work fine, as long as you find a computer powerful and irrational enough to run it.

Re:computer?

[schwaang]

I'm absolutely certain such systems would be greeted as liberators from the drudgery of all the planning we do now for these complicated military and political situations.

Re:computer?

[Cassius Corodes]

When I read stuff like this it always amazes me that people actually think such things can be predicted for any degree of usefulness. Pick up a book on chaos theory and complex systems and you will see that for large interconnected systems a small change in the starting conditions results in a completely different result. No computer could have predicted 9/11 and yet it has had a massive change in how the world works - same thing if some crazy tomorrow assassinates the president.

Even if we assume these random events not to occur you would need *all* the data to get a result approaching reality.

Re:computer?

[thrawn_aj]

Of course, you would also read that not all systems are inherently chaotic. It is by no means obvious that human society is complex enough to be called unpredictable in principle. People who tout their own "free will" should think long and hard about that and realize that simply being able to imagine a multitude of choices does not mean that each is likely to occur. Remember that a human being living in society has more in common with an electron BOUND in a crystal than a free electron. The former has several constraints while the latter is in principle unpredictable.

Readers of Asimov will know the qualitative reasons for why such things as broad socio-economic-historical trends and the actions of large groups of people can in principle be made predictable. For a system to be chaotic, it must have a large PHASE SPACE of possibilities (physical size is not always important but it is significant). What matters is the degrees of freedom and how parts of the system are coupled to other parts. Do small perturbations in the system dissipate or do they spread? Modern society has evolved into a 2-phase system where it reacts to new perturbations by simply breaking them into two possibilities - this helps relieve tensions and most people get stuck in one of the two states. This has the rather fascinating effect of re-stabilizing the system despite the introduced disturbance.

So, as the above example leads us to suspect, modern human societies are just not as complex as our egos would lead us to believe. There is strong coupling between its parts and few people stay undecided about issues - they simply get stuck orbiting one of two strong attractors in the space of possibilities and this serves to relieve any stress. In such a system of course, revolutions (in the sense of widely held beliefs changing within the lifetime of a single individual) simply cannot happen. At the worst, there might be a slow decay and unraveling of the social fabric. Barely noticeable.

Equivalent arguments apply to the "free will" of individual human beings. Humans tend to congregate in packs - behaviorally, philosophically or otherwise. This strong tribal leaning that is presumably built into our genes ensures that most behavior patterns will be statistical in nature. Indeed, the actions of an individual can be simply predicted to a first approximation by merely qualitative means even in the absence of complete information by assuming rational behavior. A better approximation can be achieved by modeling the level of rationality of the individual and assigning probabilities based on that.

While human beings may not be predictable in a strictly deductive sense, most people are (for better or for worse) rather mundane in terms of how eccentric they can be (in a way that actually affects other parts of society). This can hardly be a bad thing as the timescale of societal change must be greater than the lifetime of an individual for a society to be called "stable". If it is MUCH greater, we would call that society degenerate or decayed.

Re:computer?

[johnsonav]

No one could have predicted 9/11 or JFK's assassination. The actions of small groups of whack-jobs will always fly under the radar. Modern technology amplifies individual acts. 9/11, the Oklahoma City bombing, Columbine, all serve as examples. There are not enough computing resources in the world, to track the potential actions of each nut. But, prediction becomes less intractable when you ask the right question. You can't just say, "Okay, here is the state of things now. What's going to happen tomorrow?" Rather, "What would happen tomorrow, if the US was struck by a massive, distributed, terrorist attack (like 9/11)?" The prediction moves away from the lone, unpredictable, nuts, to the actions of the masses and well known public figures(President etc.).

I'm not saying the predictions would be good or even possible. But, the potential is there for a system that can analyze more data than a thousand RAND Corporations, and perhaps come up with better solutions.

But what I'm really worried about is the government spending so much time staring at their computer screens, waiting for the next prediction, that they miss the real action going on right now.

Re:computer?

[scottrocket]

I saw this in a Superman comic once: cosmic rays were to turn skyscrapers into rubber, & a large ship was supposed to accidentally sail up main street Metropolis.

All signs pointed to "yes".

Re:computer?

[Jame_Retief]

You are, of course, presuming that Asimov was doing more than writing good FICTION. I sincerely doubt that any program will have any noticeable success in predicting anything, regardless of the wads of cash thrown at it to make it 'better'. Remember all the computers in the classroom priorities of the last few decades? How many of you used a computer in ANY bloody classroom that did not relate directly to the class? (ie- C+ programming {more likely FORTRAN}).

Re:computer?

[Cassius Corodes]

What worries me most is that people seem to lose their common sense when it comes to computers. I can just imagine a scenario where the human intelligence gathering pointing to an event is dismissed because the "computer" didn't predict it.

Re:computer?

[hedwards]

The main variables are specificity and time frame, if you give a fuzzy enough prediction over a long enough time frame it will happen, guaranteed. The key that you are alluding to here is getting it accurate enough to be useful while in a time frame that is reasonable to work with.

There is some degree of logic in a system like this, just not in the predictive sense. A system like this is much better used to make assessments of what conditions are like. As in what outcomes are favored by a change in conditions.

It is pretty much assured that a predictive system will never be accurate enough to make a difference as an awareness of what specifically the future holds invariably changes the behavior of the system.

Re:computer?

[Anonymous Coward]

Build the irrational/randomize functions on whatever flows from Miss Teen USA 2007 - South Carolina's mouth, and buy up some Storm Worm Botnets. Problem solved.

Re:computer?

[cheater512]

Predicting especially the US Army isnt terribly difficult.
You could do it with a switch:

switch (case)
{
    default:
        fireAtWill();
}

Re:computer?

[innerweb]

We already have companies that devote most of their energy to this. We call them marketing companies. They have a pretty good idea of how the average consumer (and an even better idea by demographic) will respond to a given stimulus. They also have a pretty good idea how well a product will sell in a given market. Companies do not spend billions of dollars per year on marketing without many facts to back it up. Kind of the same issue we have with Spammers and their Chums.

I do not think the software will be used to predict individual behavior so much (though I am certain the day is coming where that will be possible rather precisely) as it is going to be used to predict group behavior. Individuals can have a butterfly effect, but masses tend to behave more predictably.

InnerWeb

Re:computer?

[thrawn_aj]

You are, of course, presuming that Asimov was doing more than writing good FICTION.

Simply trying to save time by pointing readers to a good source for a feasibility analysis of psychohistory (the statistical prediction of future trends). Science fiction may be "just fiction" but in my experience, and in the hands of an honest author (which the good Doctor indubitably was), it offers excellent first order feasibility studies of speculative sciences.

The program in the article may have limited success, but all I said in my original post was that it should, in principle, be possible to have a working psychohistory because any modern society has such strongly coupled components that individual freedom of action does little to produce broad changes in socio-economic patterns.

So, I do disagree with your statement that "any program" would fail at this. Our computing power and programming panache may be too primitive to build such a program today but I have a feeling that it can be done ... someday.

Re:computer?

[Anonymous Coward]

I agree that it is comparatively easy to predict the rational reaction to something catastrophic -- everyone runs to save their life, get food, etc. You don't need a computer program to tell you that, although a computer program might simulate the aftermath based on resources known to be available in the vicinity. However, a computer program cannot predict the event that leads to the initial conditions of the simulation. This is because it could be something completely external, like Sun radiation levels changing.

One thing that I disagree with in your post -- you keep saying "most people". Sure, most people are very predictable. But sometimes it only takes one person (or a small group) to change things completely. Those would be the most "dangerous" people and this computer program cannot predict their actions.

Re:computer?

[smellotron]

It is by no means obvious that human society is complex enough to be called unpredictable in principle... While human beings may not be predictable in a strictly deductive sense, most people are (for better or for worse) rather mundane in terms of how eccentric they can be (in a way that actually affects other parts of society).

There's still the issue of dealing with the tail end of any distribution. I don't care about the 99.999% of people who, in the aggregate, fit a model. I care about that 0.001% of people who are going to completely blow it (because, as always, "past performance does not implly future performance").

Re:computer?

[sg_oneill]

"the statistical prediction of future trends"

ugh. I'm not even going to begin to explain whats wrong with this whole damn idea.

Re:computer?

[jotok]

So, as the above example leads us to suspect, modern human societies are just not as complex as our egos would lead us to believe.

I do not think that word means what you think it means.

One person probably operates under severe constraints. But it does not follow that the behavior of society, then, is merely the aggregate of the observed behaviors, or that you can model it as a group of individuals. This is also why modeling the behavior of an ant colony is not that same as modeling one ant one hundred times: there's emergent behavior that "you don't see until you see." That is, interaction effects don't reveal themselves until you have some interaction to observe.

Also to consider is that we use different concepts and language to describe what is happening at different levels of the complexity hierarchy; how we experience and talk about individual experiences is different from how we experience and talk about group experiences, and it is not always useful to describe group phenomena as an aggregate of individuals. I explain this to my students as follows: on some level it's useful to discuss a platoon of soldiers as 30 individuals. But on another level that platoon is atomic, and it becomes less useful--exactly as it becomes irrelevant to discuss the meaning of "half a rifleman" (except when discussing a particularly gory shoot-'em-up movie).

It rather sounds like you're thinking in terms of thermodynamics (stability) rather than information theory. I always got the impression that complexity theory was over Asimov's head--it's why the Foundation series is so simple (but a good read nonetheless).

Re:computer?

[Jame_Retief]

I disagree. A chaos predictor cannot be more accurate than simple guessing, for it is attempting to predict chaotic events. Look at the political scene in the US; given the current climate everyone is attempting to say that the Democrats are going to take the election (and they might). It is just as likely that one of the Republicans will nose over the line (especially with the right third-party candidate). We know his system well and experts can only call a race with 50% accuracy, even with just two candidates in most races. The human mind is far more complex and makes far more calculations, with random associations for a computer to predict anything about what a group of people will do. I will be impressed if we see a program, ever, that hits greater than 20% reliability on predicting events or trends.

Re:computer?

[bindo]

Most of your reasoning is interesting. In fact having read Asimov I liked the way you put it.

Although... I have a simple suspect.
You have NO history or social sciences or anthropology formal education.
You like having your facts simplified so you can "reason" deeply.

Modern society has evolved into a 2-phase system where it reacts to new perturbations by simply breaking them into two possibilities - this helps relieve tensions and most people get stuck in one of the two states.

Never put your foot out of WASP society, have you?

Even if this statement would be true who says "modern societies" are important in the context discussed ? They are a minimal part of humanity and are NOT a template of where ore cultures are heading anyway....

Society IS conservative, but as a whole there is so much more happening that you point is useless....

BTW jumping to conclusions without any ability to argument seriously my personal opinion... I totally agree with you conclusion:
Of course, you would also read that not all systems are inherently chaotic. It is by no means obvious that human society is complex enough to be called unpredictable in principle. :))

Re:computer?

[OldBus]

You are, of course, presuming that Asimov was doing more than writing good FICTION.

Not only that, but in one of his later essays he discusses why the later Foundation novels look towards an alternative to psychohistory ('Gaia'). I can't remember all the details, but Asimov felt that chaos was one of the biggest reasons why it would fail (along with the problems of finding alien species from outside the Galaxy).

Re:computer?

[vegiVamp]

> You are, of course, presuming that Asimov was doing more than writing good FICTION.

Around 1865, some silly person thought he was a decent writer, and penned down 'De la terre à la lune'. Crazy, really. They should forbid that kind of thing.

Re:computer?

[chthon]

What everybody forgets is that Asimov undercut his own psycho history by introducing a random event, the Mule.

Re:computer?

[Laurence0]

Whilst I think you might be right, the example given there isn't an especially good one - part of the reason it's so hard to call the US elections is that there's so little difference between the parties. In the long run (and the short run, if you're not part of either party yourself), which sides win will have very little effect indeed.

Re:computer?

[Anonymous Coward]

Here's one prediction that is guaranteed to come true:

Government will continue to expand, both in revenue and power over the people, as history has demonstrated over and over again. The power elite at the top will continue to consolidate and centralize power, expanding the business of government and concentrating revenue and power in the hands of the few, just as they have since the dawn of organized coercion.

History shows that all governments expand in power and revenue throughout their lifetimes -- some faster than others, but governments only get bigger, never smaller. How do I know this? Because no government in history has ever significantly and permanently reduced its power or revenue through the process of democracy, and certainly not through the process of bureaucracy.

This is not something to overlook, some irrelevant detail -- this is a valuable insight into the true motives of the power elite who are in the business of government.

By the way, for all your attacks on the individual's natural human right (god-given if you prefer) to free will and self-ownership, you still haven't explained this: if man cannot be trusted with the ownership and government of his own body and mind, then how can he possibly be trusted with the ownership and government of others?

Re:computer?

[beckerist]

Isn't the only way you can accurately predict the future is to know the location, velocity and acceleration of every single particle in the universe? Isn't it also a fact at this point that the more precise you measure location, relatively the less accurate you can measure velocity as a result (and vice-versa?) PLUS, isn't the fact you are writing a computer application to predict something imply a pattern?
The article has a sub-headline: "Actions Sometimes Defy Logic." Ummmm, then how the heck is a COMPUTER going to figure it out. It can't even read a captcha, or generate a "random number" without a logical external seed (like a time variable,) let alone work illogically...

Re:computer?

[Jame_Retief]

Yes, some things do come true. Heinlein, while an excellent writer, penned stories about atomic power plants on spaceships. Not only have there not been the type of power plants he envisioned, EnviroWeenies freak at the idea of having atomics get pushed up on rockets. He did, OTH, come up with the idea of waterbeds (I think, never have researched it). Many, many things get talked about in fiction, some of them are even good ideas. Most are just the idle fantasies of writers. Like psychohistory, which smacked more of magic than of science.

Re:computer?

[Jame_Retief]

Point. Then consider the changing fortunes of countries. 100 years ago the UK was dominant in world affairs, trade, etc. Where will the US be 100 years from now? Is there an obvious successor on the world stage? China wants badly to be that successor and so does the old USSR (they may not be socialists, but Putin is no less ambitious) but how can we tell who will take it?

Re:computer?

[Anonymous Coward]

... And Lucian of Samosata in the 2nd century had his heroes carried to the moon by a waterspout.

That's just the earliest, there are many more. They pretty well all have totally impossible means of getting there, so your point is?

Re:computer?

[Lurker2288]

I think you're looking at it backwards. Ad firms don't spend millions advertising because they know how people will respond to stimuli; they spend millions precisely because there's so much uncertainty. Firms don't buy Super Bowl ad time because football fans are more susceptible to commercials--they do it because if you get your message out to such a massive audience, odds are much better you'll hit a few takers. And much of the money that's spent goes into research precisely because they want to determine, before they launch a campaign, how people will really respond to it. Even then, some things flop, and some things take off in ways nobody foresaw--do you think Geico expected the caveman to become as popular as it did?

As for predicting group behavior, what is the stock market but the grouping of millions of individual workers into neat little quotable companies? The market vastly simplifies an enormous amount of economic activity into a fairly concise set of numbers, and yet the best experts still can't do much more than guess about what the market will do at any given time. Read 'The Black Swan' by Nassim Nicholas Taleb, where he points out that every time a new model of the market is produced and its predictions are retroactively compared to reality, the model turned out to be horribly wrong. Yet we maintain this intellectual conceit that if we knew just a LITTLE bit more about the present, if we had just a BIT more computing power, we could decode everything. Silly people.

Re:computer?

[Anonymous Coward]

'--exactly as it becomes irrelevant to discuss the meaning of "half a rifleman" (except when discussing a particularly gory shoot-'em-up movie).'

Or this guy :^) (and for those who like Lego).

Re:computer?

[Lurker2288]

"Because no government in history has ever significantly and permanently reduced its power or revenue through the process of democracy, and certainly not through the process of bureaucracy" Not to be picky, but didn't the Weimar Republic vote itself out of existence? Granted, an n of 1 doesn't make for much of an argument, but...

Re:computer?

[jdjbuffalo]

While you bring up some interesting points, I'm going to have to disagree with you on a few points.

In such a system of course, revolutions (in the sense of widely held beliefs changing within the lifetime of a single individual) simply cannot happen. You are naive if you think a revolution in beliefs can't change in someone's lifetime. I'll grant you that it's less likely to happen in less than a generation (approximately every 20 years) but it still does happen. Take things like Blacks gaining the same rights as whites. Gay being accepted by society. Society as a whole doesn't necessarily change as fast as individuals, it seems to be speeding up with the advent of the Internet.

At the worst, there might be a slow decay and unraveling of the social fabric. Barely noticeable. If you look at revolutions or "key nexus points in history", they can drastically change the social fabric in a short period of time. We need to look no further than 9/11 to see how drastically one single event changed the way we as society behaved. Long held beliefs like being against Torture were suddenly acceptable to many people what wanted the government to do "anything possible" to get terrorists. I could go on an on about this but I trust my point has been made.

Re:computer?

[Lurker2288]

"There is some degree of logic in a system like this, just not in the predictive sense. A system like this is much better used to make assessments of what conditions are like. As in what outcomes are favored by a change in conditions."

Even this doesn't really hold water, though, because reality isn't like a big equation, where you can take the partial derivative of x and know how the result will change as x varies. Whether x is the population of China, the number of nuclear missiles in the US arsenal, or the odds of a given world leader being assassinated in the next year, changing x is going to change other variables in ways that aren't necessarily obvious; it's not possible, in reality, to hold all else constant. And even if you could isolate one variable, how do you determine the likelihood of a change in that variable off a given consequence? It's not as though there's any deterministic way of reaching a conclusion beyond wild-ass guessing.

Re:computer?

[vegiVamp]

My point is that, even if they often fail to accurately envision the means (and the longer before, the more inaccurate), those people still managed to see a time where those things would become possible, even if described in a manner consistent with the technological and other limits of their time.

Also note that Asimov never provided much details about exactly *what* psychohistory was (or, rather, will be) - merely that it was some advanced combination of psychology, history and statistics.

It may not always (or even often) be their explicit intention, but that's not going to stop the more learned writers from making predictions about possible future paths for technology which may turn out to be remarkably accurate. Have a look at Arthur C. Clarke's writings about the spage age in the late first half of the previous century.

Re:computer?

[scoove]

A better approximation can be achieved by modeling the level of rationality of the individual and assigning probabilities based on that.

And yet even this approach has serious problems; normal probability models fail miserably when applied to numerous areas where humans have an influence. Check out Dr. Taleb's The Black Swan: The Impact of the Highly Improbable or at a minimum, check out Dr. Taleb's website for more details.

Taleb's conjecture (which is supported by many others in the fields of behavioral finance & economics) is that models based on normal distributions (the primary model in Taleb's "mediocristan") fail when they encounter conditions often found with human behavior that seriously violate the prerequisites for such a distribution. For example, a normal distribution requires that each flip of a coin be an independent event and have no outcome bias from previous heads or tails. Unfortunately, many human events tend to violate this precondition. Winners tend to win more, and losers lose more, in human situations possibly due to herding behavior (e.g. people feel they will be safer from harm by being around a winner, thereby assigning more resources to the winner and allowing them to win even more). This condition alone damages normal distribution models and tends to yield what is called a fractal distribution (also known as a "power law" distribution).

Other conditions, such as bimodal and multimodal models, further distort the use of normal distribution approaches in forecasting events. Some suggest the economy is bimodal, having a "good economy" ruleset and a "bad economy" ruleset based on the majority perception. A friend who manages a multi-billion dollar mutual fund has conjectured that U.S. financial markets are multi-modal, with its behavior corresponding to the financial belief system of the dominant participants. E.g. the dot-com boom was driven by non-technical investors who believed in the rule of capital appreciation (and totally ignored models like dividend discount model and such for equity valuation). Following that bust and their departure from the market, those who remained in the market were mostly those educated in classical financial analysis methods, and subsequently the market tended to behave following their rules for a few years (mostly 2002-2004). When technical analysis investors tend to get active, you'll find equities of interest to the TAs tend to start following the TA rules - mostly because those who are interacting with it daily believe in them. This goes on until a different dominant force influences the mode, often due to opportunities being discovered by users of other models that are missed in the current one. This approach was also heavily practiced by George Soros, who would raid an investment target when the current mode carried it way out of line, creating unique risks (such as liquidity risks) which the current model didn't recognize. Multi-modal distributions cause normal distribution approaches to fail miserably as the modes tend to be organic and have considerable influence from exogenous events.

Then again, it's not terribly surprising that grants such as these are given. Most of our financial and risk analysts are trained in classical models and are constantly shocked when the real world doesn't behave as such. If you're interested in this kind of stuff, or find amusement in the failure of supposedly smart people to predict stuff, check out "Why Most Things Fail" by Paul Ormerod,

Re:computer?

[Anonymous Coward]

Predicting who will be elected in the next usa presidential election is indeed very difficult. However, you can say that it will most likely be a democrat or a republican (imo 99%). You can also say it will most likely be someone who has a lot of money and a lot of support (imo 90%). You can predict the last two rather well, the former is very hard to do.

Re:computer?

[foobsr]

When I read stuff like this it always amazes me that people actually think such things can be predicted for any degree of usefulness.

I doubt that those who fund actually believe that, but are in need for a proper envelope to direct the money to worthy and willing puppet researchers usable for other projects as well (I have the suspicion that there are a lot of those at the 'University of Arizona', all keen to use AI for repression, among them a Fei-Yue Wang (Chinese Academy of Sciences), especially interested in using robots in riot control — go figure).

CC.

Re:computer?

[recharged95]

Hence, law is a strange attractor, therefore in itself, civilization.

Sounds describe-able (not predictable) to me.

Re:computer?

[ArsonSmith]

Other than it shatters your preconceived notion of free will?

Re:computer?

[ArsonSmith]

what's the old saying?

"Any sufficiently good fiction is indistinguishable from fact."

Re:computer?

[WindowlessView]

everyone is attempting to say that the Democrats are going to take the election (and they might).

Hey, Democrats win elections, Republicans take them!

Re:computer?

[Jame_Retief]

everyone is attempting to say that the Democrats are going to take the election (and they might).
Hey, Democrats win elections, Republicans take them!


Think what you want. Chicago is heavy with Democrats and everyone who has EVER died in Chicago votes at least once in every election for their favorite Democrat.

Not to mention, if you paid attention in Ohio, the slant made by Diebold errors was in FAVOR of the Democrats (who threatened outright violence if they didn't get their way). In Florida the Democrats were ready to revolt and never got closer than 500 votes, even after NUMEROUS recounts, even after the Supreme court spoke up.

AND there is a particularly notorious event out in Washington state where the Democrats openly stole an election by disqualifying all mail-in ballots (primarily military). It was JUST enough to get the Dem candidate over the Repub candidate. Less than 10 votes, IIRC.

Neither side is innocent, though.

Re:computer?

[thrawn_aj]

Other conditions, such as bimodal and multimodal models, further distort the use of normal distribution approaches in forecasting events.

Sure. One could then simply use Bayesian analysis (once its machinery is developed further) and give the entire distribution as a prediction. The physical interpretation might be something like a metastable state where the system might flip between two or more states rather easily or may even be made to do so by deliberate manipulation.

As more general observation, I find it amusing that the mere statement about the limitations of free will and the possible existence of a statistical version of "humanity" made so many people rear up on their hind legs :P. Well, it's hardly the first time I've been called naive, but I would have been a little more swayed by people's arguments if their ad hominem attacks didn't move me to tears (uncontrollable chuckles actually but let's not quibble =D). And what's with all the Asimov-hatin'? :P Neither I nor he (in his writing) ever said that we had a PROCEDURE for such statistical prediction or that such a thing was even practically possible (perhaps with a lot more computing power but more importantly with newer and better analytical methods - Bayesian data analysis seems like a good candidate: I've just started learning this so I'm not sure yet). I merely stated my opinion that free will is highly overrated in terms of its ACTUAL existence in today's society. I am actually of the opinion that we DO have free will. For several reasons, most of us CHOOSE not to exercise it (teenage "rebellion" is actually nothing of the kind - it is merely a transference of control from parental authority to that of one's peers and as such is highly predictable). This could be simply due to the fact that we live in an interconnected society where we try to behave in accordance with established norms and deviations that interfere with others' lives are punished in some form or the other.

That is the only point I was trying to make - individual components when allowed to interact in sufficient numbers, tend to form conglomerations with highly predictable qualities. As a reader pointed out, I was indeed referring (partly) to thermodynamics, but more aptly to the generic phenomenon of "emergent order". As such, the broad trends CAN imo be plotted in terms of probabilities (this is done to some extent today - I don't see any fundamental reason why it can't be extended to more problems).

Anyway, I've seen very few insightful replies to my post (which itself was hardly the best possible exposition of these ideas - blame it on lack of caffeine :P or my so-called naivete) so I'll just shut up now :P (unless I see something worth replying to - could happen :P).

Well...

[427_ci_505]

Didn't see that one coming.

Re:Well...

[bencass]

They just need to hire a few psychohistorians to program it.

Re:Well...

[vertinox]

Didn't see that one coming.

I'm sure the computer did.

Re:Well...

[thewils]

I don't blame you. If it is unpredictable then by definition it can't be predicted, the best you could manage would be an educated guess. I guess they can really only predict the stuff that's just mindbogglingly difficult to predict.

If it really can predict the unpredictable, let's see it try to predict which atom in a radioactive element is going to decay next. Then I'd be impressed.

Re:Well...

[Anonymous Coward]

> > Didn't see that one coming.
>
>I'm sure the computer did.

The Computer sees everything. Trust the Computer! The Computer is your friend!

Re:Well...

[Anonymous Coward]

I almost choked on a cookie! >:(

Re:Well...

[JohnVanVliet]

as i was reading i was thinking how nice it would be if the psychohistorians from the Foundation trilogy + 3 prequels were here

Re:Well...

[Anonymous Coward]

Software predicts bugs in software.

Re:Well...

[Anomolous Cowturd]

Now they just need $2.2 quadrillion to develop a telepathic positronic robot to play god.

Re:Well...

[CryptoKiller]

dammit, I almost sprayed my morning coffee all over my machine after reading that one! :-)

Re:Well...

[AcidLacedPenguiN]

Didn't they already make a computer program that predicts the unpredictable? Isn't it called Puzzle Quest?

Re:Well...

[Anonymous Coward]

Computer says No.

bullshit flag

[Anonymous Coward]

Go ahead and predict the weather for a week. I will be impressed.

Predict it for 2 weeks, I will blow you.

You cannot predict something with so many variables that you don't understand. You certainly cannot do it regarding how people will react.

Re:bullshit flag

[Anonymous Coward]

A/S/L?

Re:bullshit flag

[Anonymous Coward]

I can predict the weather 2 weeks out. The temperature will be between -150 and +200 degrees fahrenheit. There will be between 0 and 20 feet of percipitation. Your place or mine?

Re:bullshit flag

[localman]

I once heard that if you predict the weather tomorrow will be the same as today, you'll be right more often than most meteorologists.

So lets start now on my free blowjob. How many chances do I get? ;)

Re:bullshit flag

[zeromorph]

bullshit flag

I second that.

The whole article is totally bizarre and buzzword populated begging for attention. Not only will it predict the actions of nearly every bunch of lunatics it will also "display data in graphical, 3-D and other forms that can be quickly grasped".

Please! We have a highly complex situation, with a lot of different agents and a long genesis, and literally millions of different contextual factors influencing the situation and they take all this munch and crunch it a little with fancy buzzword concepts and put it in a pie chart?

This is an insultingly brazen self-adulation.

While the software ultimately could save millions of lives,...

Ok, I changed my mind I'm gonna die laughing.

Re:bullshit flag

[networkBoy]

Yes you can.
The weather and radioactive decay are random events.
an individual is difficult to predict; large groups, however, are rather predictable.
The whole issue with using humans to predict political situations is that humans are biased.
The computer is not. Even with common polls showing the election one way ENIAC predicted the presidential election correctly for the expected loser.
-nB

Re:bullshit flag

[UnanimousCoward]

What do you look like? On second thought, who cares?

Re:bullshit flag

[Anonymous Coward]

Go ahead and predict the weather for a week. I will be impressed.

Predict it for 2 weeks, I will blow you.

Here we go: the weather, for the next two weeks, will be very sunny at 100 degrees. Blowjob please. Wait, on second thoughts, I'll donate my winnings to charity. Give it to a hobo instead.

You cannot predict something with so many variables that you don't understand.

Of course you can. I just did. Whether that prediction matches reality is another thing entirely. James Randi you are not.

Re:bullshit flag

[Anonymous Coward]

Predict it for 2 weeks, I will blow you.

You will suck mah bahls, [AC].

Re:bullshit flag

[KDR_11k]

But the weather is just large groups of gas and fluid molecules!

Re:bullshit flag

[futuramarama]

I was listening to Ray Kurzweil talk about thermodynamics, where although the system consists of enormous number of different factors operating in ways too difficult to track, the overall result could be plotted on a 2d graph. ie a bowl of soup will cool at a regular rate. I'm not sure I captured his point very well, but the idea is that even though a country's political climate is enormously complex, it may follow a trend at some level that could be plotted.

Re:bullshit flag

[DocSavage64109]

Well sure you can plot anything in two dimensions as long as you drop enough variables. You can at least use color to make your 2d chart contain additional data at the expense of colorblind people being able to differentiate some of it.

Re:bullshit flag

[Anonymous Coward]

Predict it for 2 weeks, I will blow you. That's all? Alright, two weeks from today, we're looking at a high of 120 and a low of 80. Mostly sunny, possible drizzle. That'll be one bj please.

Re:bullshit flag

[Anonymous Coward]

As one of the previous commenters said, it will take six months of calculation to predict the weather.
Luckily, the researcher

1. Predicted in advance that he was going to get the funding and bought the hardware a year ago, and

2. Predicted your cocky remark, and built up the weather predictions for the next two weeks.

3. Unzipped his pants a week ago..

Re:bullshit flag

[JesterXXV]

Agreed.

"Predict the unpredictable" does not compute.

Re:bullshit flag

[Anonymous Coward]

Predict it for 2 weeks, I will blow you.

After reading that post, I thought I misread the title to be "bullshit fag"

You don't need software for that

[fedxone-v86]

Just flip a coin and / or ask the stars.

Re:You don't need software for that

[pilgrim23]

Speaking as someone who worked his way through college reading Taort cards in local coffeehouses (I am not kidding, I really did).... I don't believe a word of it..

Re:You don't need software for that

[I'm New Around Here]

Or ask people that aren't certified behaviorists or criminal profilers. Ask the kids who were picked on in school, whose embarassment and anger could only be contained behind a pleasant facade. Here's a case from my view.

In 1993, terrorists tried to destroy the World Trade Center by exploding a large truck-bomb in the basement parking garage. It did a lot of damage, but nothing severe. The government reacted by not allowing trucks in the parking garage anymore. To assure this, several large cement pylons and other traffic barriers were placed out front, and only cars and light trucks/SUVs were allowed through. Seemed like a good plan. Sorta.

My first thought after reading about these cement barriers was how could you drive trucks through these barriers to bomb the buildings again. You can't, they are too thick and reinforced. You have to go _over_ the barriers somehow. Can you drive the truck over the barrier? Not very likely. So how would you get the truck over the barrier and into the building? Put two wings on the damn thing and fly it in. In other words, rent a cargo plane, fill it full of explosives, and fly it into the buildings. In reality, the bastards were even more twisted than I would have been.

So now, would the computer be able to predict that outcome? Or could that predictions only come from a twisted brain that spent several years wanting to kill many people? Because that is exactly what we are facing. Unpredictable scenarios are only for people with no _personal issues_.

Re:You don't need software for that

[Anonymous Coward]

it doesnt matter. predict the cargo plane and have sams stationed around the wtc and voila -- problem solved.

Re:You don't need software for that

[thedave]

Just ask slashdot.

Re:You don't need software for that

[I'm New Around Here]

It does matter. Predicting the cargo plane is not an easy thing to do apparently. Or at least not back in 1993-2001. And not for the suit-wearing elite. Do you remember how Condoleeza Rice commented on the attack?

From http://www.cnn.com/2004/ALLPOLITICS/04/08/rice.transcript/

And I said, at one point, that this was a historical memo, that it was -- it was not based on new threat information. And I said, "No one could have imagined them taking a plane, slamming it into the Pentagon" -- I'm paraphrasing now -- "into the World Trade Center, using planes as a missile."

Whether or not some people can think of these horrible actions, the ones who are supposed to protect us can't. They have to pay millions for a computer to think of this shit for them.

Re:You don't need software for that

[fractoid]

OK, then try an evolution of it. Load a cargo plane with a kinetic missile, calculate the ballistics, and drop the thing on your target. Consider A 100-tonne steel kinetic missile dropped from 11km up, from a plane moving at 800km/h. These figures (cargo weight, altitude, velocity) are easily achievable by a 747 airliner.

According to the back of an envelope, this missile will strike the target at around 1600km/h vertically, giving an overall velocity of 1800km/h, 47 seconds after being released. The missile will travel around 10km laterally, allowing it to be released well away from the target. Energy released will be on the order of 12.5 GJ, which if Wiki and I are correct, is around the equivalent of 3 tonnes of TNT, but the majority of the destructive power will come from the pure mechanical pulverization of the target.

Re:You don't need software for that

[Magada]

How do you propose to aim this?

Re:You don't need software for that

[fractoid]

A combination of high-precision GPS, compass, laser altimeter and some computer code with lots of sin() and cos() and goto(). :P It'd be moderately expensive to develop such a system, but much cheaper than buying a 747, which you also have to do. On the plus side, both the code and the 747 would be reusable (well, assuming the sidewinders miss the 747, but that's another story. :P )

In practical terms you'd probably have a small self-guiding system on the missile, something along the lines of the systems used on laser guided bombs, although obviously with somewhat larger wing surfaces. That should be enough to nudge the whole thing in the right direction, given that it was roughly on target.

It's A TRAP!

[lennier]

The Asymmetric Threat Response and Analysis Project, known as ATRAP, is a massively complex set of computer algorithms (mathematical procedures) that sift through millions of pieces of data.

They come right out and say it...

Re:It's A TRAP!

[wickedskaman]

Can this parent be modded above five. That was a thing of unbridled beauty...

Re:It's A TRAP!

[hazem]

Well, it's natural considering its predecessor was the Complex Response Analysis Project, but after huge overruns, that project had to be flushed.

Ridiculous

[Reality Master 101]

Apparently there are those that have forgotten the old computer law of "Garbage In, Garbage Out". Even if we had a perfect model to predict these sort of things, we don't have any way of supplying the required data to model the prediction. What's the computer going to do, go undercover in secret groups? Read the web sites? Listen to radio chatter and analyze their conversations?

Maybe someday when we have a real science of A.I. something like this might be possible, but all it shows is that this university professor will happily take government money for delivering absolutely nothing.

Re:Ridiculous

[paulthomas]

Moreover, wouldn't the existence of such a prescient system influence outcomes?

Re:Ridiculous

[ChrisMounce]

Maybe someday when we have a real science of A.I. something like this might be possible, but all it shows is that this university professor will happily take government money for delivering absolutely nothing.

He has already perfected the software and is using it to game the grant system.

Re:Ridiculous

[Jeremi]

Apparently there are those that have forgotten the old computer law of "Garbage In, Garbage Out" [...] all it shows is that this university professor will happily take government money for delivering absolutely nothing.

Unless, of course, garbage is what they are after. Last time it was "Curveball" that gave them the necessary disinformation to justify a war; next time they won't even need to bother with informants, they'll just look to their computer program to tell them an invasion is necessary. Accuracy would only get in the way of the political goals anyway.

Re:Ridiculous

[56]

Even if this were to work (which is by no means a given), people will begin to take its predictions into account in order to do the unexpected. Truly a preposterous endeavor.

Re:Ridiculous

[syrinje]

While a bad use of tax-payer money, delivering nothing is the benign outcome of this whole fiasco - if they did deliver a half-assed electronic magic 8-ball of sorts. That would be a Really Bad Thing (TM). I shudder to think of the consequences of a government making choices on use of peoples lives and deployment of deadly force based on this.

OF course, no doubt, by some weird law of self-fulfilling prophecies - it will predict the end of the world and it shall be so!

Re:Ridiculous

[jagnich]

One way around the "garbage in" problem is to move one level up the intelligence ladder from artificial to human, because humans can filter out a lot of the noise. But What we need is a method of aggregating dispersed human knowledge in an intelligent way -- for example, prediction markets have proven accurate at forecasting elections and geopolitical events. A few startups (e.g. www.predictify.com) are experimenting with the 'wisdom of crowds' model to predict real-world events. It will be interesting to see whether any of these new ideas work.

Re:Ridiculous

[dbIII]

Good point. The Brits lived in India, their kids grew up there, learned the language and knew what was going on. With a more more recent empire it's fly in, fly out, rely on whatever a translator says and try to work out from photos whatever is going on. There is also the increasingly common problem of things like fake intelligence scripted by a PR company getting mixed in with the real stuff (sixteen thousand plus WMD sites remember) and all the futile horror of people that think they will be murdered if they don't saying whatever they think the torturer wants to hear.

Re:Ridiculous

[chthon]

Isn't that what the NSA and the CIA are for ?

Re:Ridiculous

[halcyon1234]

But what about the unknown unknowns? How will it know to know them?

It will always answer:

[Anonymous Coward]

"Insufficient data for a meaningful answer."

Re:It will always answer:

[fractoid]

Well, not *always*. But right until the end, yes it will. :D

come on now

[flynt]

'Predict' has a specific meaning in statistics and machine learning. It definitely does *not* mean accurately predicting outcomes in every situation. Not to belittle this group's work, because it is no doubt important and complicated, but it is not going to magically 'predict the unpredictable'.

Re:come on now

[Prof.Phreak]

Indeed. One can view playing the lottery as `predicting'. As with any other `predicting', there's a probability of being wrong...

Re:come on now

[Alpha830RulZ]

That's a good comment. However, a lot of machine learning work looks like predicting the unpredictable to the uninitiate. I have no doubt that the problem is approachable theoretically, or that it wouldn't be useful - if-, and that's a big if in my mind, it is practically achievable.

As you note, the probable output wouldn't be in the form of "this is going to happen", but rather "this outcome is more likely than that outcome", to the degree of some percentage probability. That is certainly a useful form of information even if it isn't absolute. It would be extremely useful in elections, for example.

I don't see the problem as intractable, it's a fairly logical extension of current data mining and predictive modeling techniques that are well described in literature. See http://gnomo.fe.up.pt/~nnig/papers/boo_bag.pdf/ this, for example. The problem is how to define the model training data, and to do so in a way that can be achieved in the context of when you are trying to use the output. It would boil down to trying to find a way to express the problem in a modelable fashion, and summing the outputs into some useful phraseology. It's quite easy to develop models that predict behavior, the question is more along the lines of how to describe and measure the behavior you are trying to model.

And anyone that understands enough of what I just said enough to realize how ignorant I am that would like a good job in Seattle, please contact me. We have work in this space.

Reason

[HTH NE1]

From "Dirk Gently's Holistic Detective Agency," by Douglas Adams,
New York: Simon and Schuster, 1987.

"Well," he said, "it's to do with the project which first made the software incarnation of the company profitable. It was called Reason, and in its own way it was sensational."

"What was it?"

"Well, it was a kind of back-to-front program. It's funny how many of the best ideas are just an old idea back-to-front. You see, there have already been several programs written that help you make decisions by properly ordering and analysing all the relevant facts.... The drawback with these is that the decision which all the properly ordered and analyzed facts point to is not necessarily the one you want.

"... Gordon's great insight was to design a program which allowed you to specify in advance what decision you wished it to reach, and only then to give it all the facts. The program's task, ... was simply to construct a plausible series of logical-sounding steps to connect the premises with the conclusion." ....

"Heavens. And did the program sell very well?"

"No, we never sold a single copy.... The entire project was bought up, lock, stock, and barrel, by the Pentagon. The deal put WayForward on a very sound financial foundation. Its moral foundation, on the other hand, is not something I would want to trust my weight to. I've recently been analyzing a lot of the arguments put forward in favor of the Star Wars project, and if you know what you're looking for, the pattern of the algorithms is very clear.

"So much so, in fact, that looking at Pentagon policies over the last couple of years I think I can be fairly sure that the US Navy is using version 2.00 of the program, while the Air Force for some reason only has the beta-test version of 1.5. Odd, that."

Re:Reason

[Ardipithecus]

This is the BS-A-Tron! Hundreds of units are in use at all levels in government and business.

I'll buy it

[Anonymous Coward]

Software to counteract Murphy's Law. What could go wrong?

Re:I'll buy it

[KudyardRipling]

[facetious]

That could threaten the very existence of the universe. Nobody messes with Murphy's Law. It's the left hand of the Al-mighty. Would you want Kim Jong Il's toy to not fail when it is on its way to its target (i.e. you home town)?

[/facetious]

Failure isn't always bad and success isn't always good.

Sunny and mild

[Anonymous Coward]

Okay, sunny and mild. Now you hold up your end of the deal.

What?! You said nothing about accuracy...

Re:Sunny and mild

[Anonymous Coward]

You win. You get a free blow.

Come on over and collect.

Perhaps you can predict where I am now (with accuracy)

Re:Sunny and mild

[UncleTogie]

Perhaps you can predict where I am now (with accuracy)

In front of your computer. Duh.

Now get over here and blow us.

{I couldn't help it...Resistance was futile.}

Re:Sunny and mild

[Kingrames]

Just in case he's wrong, I'll say you're in front of your nintendo DS, posting from there. like I can.

How about predicting the predictable first...

[mikeasu]

Hopefully, it'll bring their football predictions more in line with reality...as an ASU fan living in Tucson, I had to put up with months of "It isn't a matter of IF we go to a bowl game this year, but WHICH bowl game we'll go to!" I've been laughing all season, as bad as the wildcats are doing and how well ASU is doing... DISCLAIMER...I realize ASU has the roug part of their schedule ahead...

Re:How about predicting the predictable first...

[servognome]

just wait till basketball season you can't spell Arizona State without N.I.T. :D

Re:How about predicting the predictable first...

[mikeasu]

Bah, I'd mod that irrelevant if I could...basketball...meh DISCLAIMER...Since U of A typically has a better basketball program then ASU, I usually blatantly ignore basketball... ;-)

If it is unpredictable...

[DragonWriter]

...the program will still fail to predict it. By definition.

The article (as would be unsurprising even from the professional press, and is less surprising from what seems to be a school newspaper of the school employing the professor getting the grant) seems to be a very uncritical regurgitation of an extraordinarily puffed-up press release that seems to suggest that the professor has gotten a grant to develop something that already exist and presently has the capacities sought by the grant. Sometimes. Maybe. Really, the shifting use of verb tenses gave me a kind of mental whiplash trying to read it.

Also, I think that while this may be useful, the danger of overreliance on a system where quite literally no one using it understands how factors are really being used to generate outcome predictions are immense; if you get something that works well predictively at all, it will likely be prone to fail wildly if any of the many factors it is adapted to based on the historical data used to train it shift. Unfortunately, it is quite likely that the particular sensitivities will be opaque, and thus no one is likely to know when it is likely to fail. This is rather distinct from conventional analysis which, even though it may fail in many circumstances, where it is rigorous analysis and not just guesswork to start with, its assumptions are transparent and its weaknesses and vulnerabilities in application to particular situations can also be evaluated.

Re:If it is unpredictable...

[orkysoft]

...the program will still fail to predict it. By definition.

But it's magic! It's a computer program, which is magic to most people.

We will finally know...

[Sunshinerat]

..when Duke Nukem Forever will be released.

And to be honest, this alone is worth the expense.

Re:We will finally know...

[El_Oscuro]

But will it run on Linux?

Re:We will finally know...

[Ruie]

We will finally know... ..when Duke Nukem Forever will be released.

Remember the proof that one cannot decide in advance whether a given algorithm will finish in finite time ? There was a reason they put "Forever" in the title..

If the prediction is wrong, whose fault is it?

[z-j-y]

politicians?

Re:If the prediction is wrong, whose fault is it?

[Aetuneo]

Clearly, if it is completely wrong all the time, it is simply predicting the unpredictable for an alternate universe. We must then prepare to invade this alternate universe to allow the computer to predict what we are doing.

My prediction:

[Anonymous Coward]

Shit happens.

linear limitations...

[DawnArdent]

I'd rather invest in training mentats.

already done

[phrostie]

someone predicted this, i'm sure of it

Trantor

[danilo.moret]

You just need to find one single planetary system complex enough, some basic axioms, a lot of spare mathematicians and Hari Seldon to come up with a solution for predicting the unpredictable, as long as the unpredictable isn't the Mule.

Re:Trantor

[Urza9814]

No, you just need a colony of mind control capable citizens hidden from the rest of society (not hard when they have mind control) to ensure things always go the way you want.

Re:Trantor

[Anonymous Coward]

This deserves an infinite score.

P.S. Did you know that Psychohistory is a real discipline? It didn't even borrow its name from Asimov.

P.P.S I'm actually taking a break from re-reading Prelude to Foundation to reply to this.

I wrote a program to predict lottery results

[netsavior]

It was accurate an astonishing 1 in every 165 million times. I used a brute force decryption strategy.

ATRAP?

[cyberzephyr]

Hmmm, And "SKYNET" became self-aware when?

pork barrel money pit

[Anonymous Coward]

This really sounds like a huge pork barrel money pit. I bet you could dump a billion dollars into this project and not get any usable results.

Can it predict the weather six months out?

[Anonymous Coward]

I'd like to know when I should take my Spring vacation to the beach.

Re:Can it predict the weather six months out?

[onemorehour]

Can it predict the weather six months out?

Of course it can--did you even read the link?

It will just take about six months to calculate the result.

Re:Can it predict the weather six months out?

[fractoid]

Hobbes: Well look at that. It's two minutes later than when we started.
Calvin: We did it! The time machine worked! We travelled through time!

Re:Can it predict the weather six months out?

[zippthorne]

I suggest sometime after the equinox, but before the solstice.

They did this in wargames and I hope that this is.

[Joe The Dragon]

They did this in war games and I hope that this is not hooked up to any missile launch systems and they better also have tic tack toe on there as well with a mode where the system can play it self.

The only winning move is not to play how about a good game of chess?

But...

[ePlus]

Doesn't real-time software already do this?

What happened...?

[suv4x4]

To that magical software that was supposed to guess if a movie will be a box-office homerun or not. It was supposed to turn the industry around and make poorly performing movies part of the past.

Well, so much for this one as well.

Re:What happened...?

[Dunbal]

To that magical software that was supposed to guess if a movie will be a box-office homerun or not. It was supposed to turn the industry around and make poorly performing movies part of the past.

      Seems to me like they are using it. The software keeps predicting more income by making sequels than original movies...

Re:What happened...?

[caffeinemessiah]

There have been attempts to do that in multiple fields, including Marketing "Science". However, if you're going to bet on a system that can predict box office sales, I'd put my money on a computer science approach. Here's an interesting paper from 2006: Predicting movie sales from blogger sentiment.

Asimov anyone?

[Ristol]

Should we clook forward to a Rozenblit plan sometime in the future?

GL style

[Anonymous Coward]

Predict the unpredictable
Compute the uncomputable
Row, row, fight the powar!

Meh...

[mark-t]

Let's see them predict the outcome of a series of discrete random events with a statistically significant greater success rate than the mathematical probabilities of the events would suggest.

Fix weather prediction first

[EmbeddedJanitor]

I'll remain a skeptic until I see reliable weather prediction at least.

Re:Fix weather prediction first

[HeyBob!]

The next few months in North America will be colder than the last few months.

I can save them a bunch of money...

[msauve]

I'll sell them a dartboard for only $1 million!

Isaac Asimov already thought of this

[the_humeister]

And he called it "psycho history." The way he describes it, it almost sounds plausible as long as you have trillions of people, and they don't know what your predictions are.

Re:Isaac Asimov already thought of this

[Ristol]

I always thought that was funny. If his prediction models were so good couldn't he just factor in the fact that people were aware of the results?

Re:Isaac Asimov already thought of this

[Jeremi]

I always thought that was funny. If his prediction models were so good couldn't he just factor in the fact that people were aware of the results?

That would give you an infinite recursion, no?

i.e. then you'd have to factor in the fact that people were aware that you had factored in the results, and then factor in the fact that they were aware you had factored in the fact that people were aware you had factored in the results.... and so on until your head explodes.

Re:Isaac Asimov already thought of this

[Dun Malg]

I always thought that was funny. If his prediction models were so good couldn't he just factor in the fact that people were aware of the results? It was a plot device, not real science. He could make up any limitations he wanted.

Re:Isaac Asimov already thought of this

[Colin Smith]

We call it "economics".

Can it predict governmental trends?

[SystemFault]

Can the system predict governmental trends? I'm thinking of something like when a somewhat peaceful, balanced budget democratic republic passes some multidimensional cusp and becomes enamored of gross deficit spending, preemptive military strikes, illegal detainment, torture, and fascism in general? It sure would be nice to see this coming a few years in advance.

Re:Can it predict governmental trends?

[mh1997]

Can the system predict governmental trends? I'm thinking of something like when a somewhat peaceful, balanced budget democratic republic passes some multidimensional cusp and becomes enamored of gross deficit spending, preemptive military strikes, illegal detainment, torture, and fascism in general? It sure would be nice to see this coming a few years in advance.

Why, 1945 is long gone (Korean war in 1950)and we've (USA) only had about 5 or 6 years of surplus since then - and the surplus only occured because both parties of congress were controlled by one party and the presidency by the other - cancelling out each other's overspending ways.

Re:Can it predict governmental trends?

[servognome]

Can the system predict governmental trends? I'm thinking of something like when a somewhat peaceful, balanced budget democratic republic passes some multidimensional cusp and becomes enamored of gross deficit spending, preemptive military strikes, illegal detainment, torture, and fascism in general? It sure would be nice to see this coming a few years in advance

Given the first state never existed it isn't really a meaningful thing to predict.

Chaos Theory

[orkysoft]

What it could do is run a simulation of the world in faster than real time, and use the result as a prediction. But then you get into chaos theory territory. The real world is very complex, and (currently at least) impossible to simulate precisely. Chaos theory says that small inaccuracies will totally screw up the simulation in the long run. Just look at long-term weather forecasts. This project seems to be nothing but a long-term weather forecast for geopolitics. I remember a similar project being underway, but can't remember a name or site. I will remain skeptical about this.

Your response is sooooo...

[WallaceAndGromit]

...predictable.

How hard could it be...

[Genda]

You create a really nice UI for a magic '8' Ball simulator... and call it a day before the brewskies warm up!

Didn't People Learn Anything from The Matrix?

[CodeBuster]

"Did you know that the first Matrix was designed to be a perfect human world? Where none suffered, where everyone would be happy. It was a disaster. No one would accept the program. Entire crops were lost. Some believed we lacked the programming language to describe your perfect world. But I believe that, as a species, human beings define their reality through suffering and misery. The perfect world was a dream that your primitive cerebrum kept trying to wake up from. Which is why the Matrix was redesigned to this, the peak of your civilization. I say your civilization because as soon as we started thinking for you it really became our civilization which is of course what this is all about..."

-- Agent Smith

Re:Didn't People Learn Anything from The Matrix?

[Boronx]

I've never met anyone who went crazy from lack of suffering.

Re:Didn't People Learn Anything from The Matrix?

[badzilla]

As a parent you try very hard to provide a great environment for your kids that gives them everything they could possibly need. Sometimes though you start to feel this is actually just turning them into spoilt brats and wonder if they would actually grow into better people if things were a little tougher for them. Not really the kind of thing you can experiment with though so you just have to cross fingers and hope!

Re:Didn't People Learn Anything from The Matrix?

[Anonymous Coward]

Giving them everything they could possibly need never makes them spoiled.
It's giving them everything they could possibly WANT that makes them spoiled.

Usefulness?

[martin_henry]

Let's see it predict spammers' & trolls' posts...

Re:Usefulness?

[Surt]

What sort of challenge is that? Seriously, has a spammer/troll ever posted anything even remotely surprising?

Re:Usefulness?

[martin_henry]

I consider eating human feces pretty surprising...but mostly just disturbing.

Read "Minority Report" By P.K. Dick.

[khasim]

The short story, not the movie. They're completely different. It covers exactly this situation.

Precog #1 sees the future.

Precog #2 sees the future that happens when you know what Precog #1 saw.

Precog #3 sees the future that happens when you know what Precog #2 saw which was the future that Precog #1 saw.

Re:Read "Minority Report" By P.K. Dick.

[paulthomas]

Thanks, I'll pick it up. I've seen the movie, and I've heard really good things about Philip K Dick around these parts.

Re:Read "Minority Report" By P.K. Dick.

[Alsee]

Wait.... you mean the movie was completely different than the book?

That's really weird.

-

Re:Read "Minority Report" By P.K. Dick.

[feepness]

Precog #3 sees the future that happens when you know what Precog #2 saw which was the future that Precog #1 saw. Yeah, but does Charlie still die underwater then?

Re:Read "Minority Report" By P.K. Dick.

[Anonymous Coward]

You should see the movie "Next" and then the P. K. Dick book it was based on, "The Golden Man." Utterly different, I don't know why they claim one was even based on the other.

Re:Read "Minority Report" By P.K. Dick.

[MrMr]

Pfff..
I knew you were going to post that.

Re:Read "Minority Report" By P.K. Dick.

[apoc.famine]

Which also leads to an important point:

If your software predicts a terrorist attack at a specific location, you will increase security at that location. If the terrorists see this increased security, they may either not attack, or they might choose another location.

If they choose another location, your prediction didn't really help, unless they chose a far less important location to attack.

If they choose not to attack, then all is well. Until this happens too many times in a row, and either the terrorists go completely random, or the general populace complains that you're not doing anything but scaring people with imaginary attacks.

Just because you can predict what will happen doesn't mean you've solved a useful problem. It's possible that you've just wasted a lot of time and energy making the whole thing more complicated.

"No One Could Have Forseen!!"

[Anonymous Coward]

Wow, something like this could have really helped out the Bush administration... seeing as how none of them were able to (supposedly) forsee anything which happened for the past six years.

It's kind of amazing how much money they made off their failures, isn't it?

The Chinese plan...

[meburke]

Chinese researchers today announced $10.2 million (USD) funding for a system to predict the outcomes of unpredictable outcomes predicted and influenced by US ATRAP computing, with the goal of further influencing the the outcomes to produce a balance-of-trade advantage for China and producing a complete domination of Taiwan...

Re:The Chinese plan...

[Anonymous Coward]

Well, comparing how well they're doing, and how badly we have messed up, I think their programmers must be a lot better.

This would be laughable if it wasn't so frightening. Do we really think we can run our foreign policy off a machine? We should go by historical results, and build a lot of British Public Schools, which produced the heros who shaped the British Empire. Read about Brooke, the 'White Rajah of Sarawack', Clive of India, 'Chinese' Gordon or Laurence of Arabia.

Prescient

[SpaceLifeForm]

Apparently, such a program is still in use.

Wrong title: Will fleece the gullible

[WillAffleckUW]

That would be a better title for the article.

In a perfect world, one could get enough data points to do such a thing.

We don't live in that world.

Infiite Improbability...

[Genda]

It's based on the "Infinite Improbability Engine (tm)", a device no galactic hitchhiker should be without...

You simply analyze the current level of improbability, determine outcomes of like chance, and match significant variables.

It also makes a great "What if analyzer". At 15 to 234682894645 Hillary Clinton becomes the next Pope, at 73 to 23456516025806291678675351675702386 Pigs do fly, at 8 to 65416944165465205141982578424752139841454586232211 Hell freezes completely over, and at 11 to 2154563256986558789995211230320012564546006567896233541056489521889662112200 George W. Bush makes perfect sense.

Happy computing!

One step closer?

[Tribbin]

One step closer to understanding women?

Finally

[Anonymous Coward]

a computer we can put in charge of Gundam!

Rozenblit's work on coevolution/genetic algorithms

[FleaPlus]

For anybody who wants a little more info than is present in the popular-press summary, here's a couple of conference papers from Rozenblit's group on using coevolution and genetic algorithms to analyze/visualize military scenarios. I think they might require institutional subscriptions to see the full PDF, but I've pasted the abstracts below.

A coevolutionary approach to course of action generation and visualization in multi-sided conflicts

The current state of military operations includes many stability and support (SASO), multi-sided conflicts. The research presented in this paper attempts to address this complex environment by creating a SASO simulation, coevolutionary generation of courses-of-actions (COAs) for each side, and visualization tools for analysis of the resulting COAs. The SASO simulation is significantly different from previous systems because it incorporates non-conventional warfare units such as terrorists and media. The coevolution algorithm is different because it allows all sides of the conflict to evolve their COAs. The visualization tools are important because SASO doctrine is not as well developed as conventional warfare doctrine. Therefore, visual analysis and understanding of a system that is not well defined provides insight for future modeling and verification.

Modeling and simulation of stability and support operations (SASO)

Stability and support operations (SASO) are becoming increasingly important in modern military operations. Conflicts are no longer comprised solely of two opposing sides engaged in combat on an open battlefield. Instead, they are more likely to involve groups sharing various alliances and relationships each pursuing a range of different goals. The Sheherazade SASO wargaming engine presented here: a) incorporates subjective criteria for scoring course of action (COA) success such as the animosity between factions and attitudes of locales, b) uses nontraditional units such as refugees, media and information operators, and c) employs a coevolutionary genetic algorithm in modeling the dynamics of the complex multisided simulation for generating COAs. This paper outlines our approach towards the development of a wargaming model that handles the more complex and computationally demanding arena of SASO.

Don't we already have a Basic Law covering this ?

[Eth1csGrad1ent]

You don't need a top end computer, or fancy algorithms... all you need to do is follow this basic principle. Might even have helped in Iraq and Hurricane Katrina, had we followed it...

I think GW needs this....

[Anonymous Coward]

Military predictions? -- Cewl.

Will it predict that if you invade a foreign country with no post invasion support plan that your administration will look like a bunch of idiots?

Absolute Idiocy

[alexborges]

I know im preaching to the converted. But doesnt this seem to you, as it does to me, like a blatant and outright liar is somwhere in the deal?

Even entretaining the idea of the existance of a software capable of this, it would sure cost a bit more than 2 million bucks to run the damned thing, let alone develop it.

Been there, done that?

[yams69]

From http://www.cs.rice.edu/~devika/projects/stoll.html:

"Events, Patterns, and Analysis: Forecasting International Conflict in the Twenty-First Century
[...]
Project Goals:

We believe that the proliferation of news in electronic form as well as a series of advances in information extraction, data mining, statistical machine learning and stochastic modeling have made it possible to predict the outbreak of a serious international conflict by analyzing event data extracted from a multitude of sources over an extended period of time. The goal of our project is to develop techniques to construct extensive event data sets and models necessary to make such predictions. We hope to be able to predict the onset of serious international conflicts four to eight weeks in advance. Specifically, the goals of our research are:

* To design information extraction techniques and build events data sets for use by the entire scientific community.
* To use these events data and develop the algorithmic base for making predictions about the onset of serious conflict.
* To construct explanatory models in the form of dynamic Bayesian networks, building on the existing findings from the scientific study of international relations.

Timely warning of the outbreak of serious conflict can be a key element in conflict resolution. Early warning can provide the time for state and non-state actors to intervene and prevent the outbreak. Thus, we feel our work can be of potential value to the conflict resolution process, even though the focus of our research is predicting the outbreak and evolution of conflict."

This project was funded 2003-2006 by the NSF.

Cease and Desist Order

[MrCopilot]

Professor Jerzy Rozenblit,

Be advised the Foundation has Patents covering the areas of study and interest.

H. Seldon

Re:Cease and Desist Order

[The_mad_linguist]

Mister Hari Seldon,

I knew you'd say that.

J. Rozenblit

Koolaid anyone?

[Anonymous Coward]

well it's good to see that someone is drinking their koolaid

Jason
http://gravityswitch.com/

When something's unique enough...

[Kjella]

...using a computer model is like driving by looking out the rear view mirror. I bet after the first major miss, they'll claim to have added that to the model. Then the next big thing will add five new factors and make three others irrelevant. Computers can't predict what they don't know what means. I'd much rather take a well reasoned human analysis over that unique situation than trying to find patterns that are spurious at best and plain out wrong at worst.

I for one...

[tjstork]

Welcome the flim flam predicting the unpredictable coding in a bunch of random number generating overlords.

What a waste of 2 million bucks.

Re:I for one...

[Phil06]

A Beowolf Clusterfuck

Re:I for one...

[jt2377]

It's better to spend $2.2 million in Unveristy doing research than to waste billion on "War for Oil". There are worst govt. project out there than this one. At least, the student at UA can benefit from this project and not dying in the desert.

Re:I for one...

[tjstork]

Well, there's 20 trillion dollars of oil in Iraq.

Re:I for one...

[jt2377]

Iraq Oil does not belong to American. Those trillion are blood money. How many more people have to die for American to get their blood money??? How many American soliders have to die? How many innocent Iraqis have to die?

Re:I for one...

[tjstork]

many more people have to die for American to get their blood money??? How many American soliders have to die? How many innocent Iraqis have to die?

Nobody has to die. All we need is for Iraqis to stop killing each other, then, we can leave, with a good democratic government in place.

Re:I for one...

[jt2377]

and who created this mess? who went to war on empty Weapon of Mass destruction lie? 90% of mess in Middle east are created by American just look at the history.

You don't get it

[passthecrackpipe]

It isn't about actually being able to predict anything useful. Think of it like this. As a "World Leader" [sic], how much would you spend on the Ultimate Cop-Out(tm)? yeah a few million is a *bargain* for what this thing can do. None of the people involved in this project are actually interested in the predictions. What they are interested in is that the *next* time they have a royal screw up, they can say: "well, its unfortunate this happened, but you see, we have really smart supercomputer. It has 3-D and stuff. And it tells us what is most likely to happen. This wasn't on the list. We only have limited resources, and this is the best way to focus those resources where they are most likely to be doing us good".

Its the ultimate repudiation. As far as I can predict, they will spend lots and lots more money on this, get some buddies in on the gravy train somewhere to boot, and they still got themselves a bargain.

We need real leaders, not computers

[Pedrito]

"...and the aftermath of natural disasters, such as Katrina."

Dealing with the aftermath of Katrina wasn't a matter of applying rocket science. It was simply a matter of simple logistics and a government that gives a shit about people. Unfortunately, the U.S. government has shown time and again under this administration that it could care less for the lives of its citizens, let alone the citizens of other countries. These problems can't be fixed by software. They can only be fixed by real leadership, something the people of the U.S. haven't shown much interest in electing...

It doesn't take software to predict that going into Iraq was a huge mistake. Just ask Chaney circa 1994. He knew it would be a major mistake, and he wasn't the only one. A lot of us were yelling and screaming to stop it before it started...

Software can't predict the future nor can they predict what stupid leaders will do. On Sept 10th, could anyone (or more importantly, any software) predict what things would be like in this country today? Even remotely? The war in Iraq, a country completely disconnected from 9/11. Guantanamo, spying on our citizens and other erosions of liberty... I doubt it. A single event and the responses by inept leadership led to a variety of disasters that nothing and nobody could have predicted.

Re:We need real leaders, not computers

[Oldav]

On the 11/9, I wept not only for the tradgedy of that day, but for the inevitable homicidal overreaction that would and did come from the US, "somone must be punished, they are someone so they should be punished" If you didnt see this coming, you have no vision of the future at all. "On Sept 10th, could anyone (or more importantly, any software) predict what things would be like in this country today?"

Re:We need real leaders, not computers

[Anonymous Coward]

The growing trend of corporate fascism in USA has been documented since the early 90's mind you.

Our mind and body is much better than any computer at calculating the unpredictable, yes.

We have been warning of this for 20 years, but the problem is when people listen with their pockets.

Re:We need real leaders, not computers

[Anonymous Coward]

The bigger a government gets, the more individuals are treated like cattle rather than unique thinking human beings, and the less the power elite cares about "the people" rather than their own budget and level of power.

The US government of today dwarfs the US government of only 100, let alone 200 years ago, both in revenue per population and power over the people. In all probability, the US government of 10 years down the road will be bigger than the US government of today, both in revenue and power over the people.

under this administration

Look a bit more closely and you'll realize that this administration is only the latest in a long, long line of power-hungry crooks whose main objective is to consolidate and centralize power, expanding the business of government at the expense of the individual. In fact, I'll bet my house that the next administration will follow in those footsteps -- maybe not exactly, but at the end of their term, government will be bigger, more centralized, more powerful, and more lucrative than it was before.

Re:We need real leaders, not computers

[AP31R0N]

My mods point always expire just before i have need of them. Ah well....

Went poorly != Mistake.

You can do the right thing, and still have it go all wrong. Somethings are just beyond your control. But that doesn't mean that you do nothing for fear of what could go wrong because of what you don't know or can't control. If some guy is beating his wife and you do nothing to stop it, you're in the wrong, you're a coward and a lousy excuse for a human being. If you try to help her, he might kick your ass too, or maybe he'll kill her. You can't control what he'll do, you're not responsible for how he responds, only for what you do or do not do. Had outside forces stayed out of it, or more of our allies stood beside us instead of voting their wallets to protect Saddam and his band of thugs, things would be very different there. Your political loyalties are blinding you to the big picture. Your sense of right and wrong is blurred by who happens to be president today.

Iraq wasn't about 9/11 specifically or terrorism in general. That was bad marketing on Bush's part. It was only ever about Saddam and the sanctions that were about to end. Leaving Saddam in power with billions in oil revenue to rebuild his military, resume WMD development and threaten his neighbors. As for the lie about him being contained, he was only contained in the sense that he was only killing his own people, he'd resume killing his neighbors as soon as the sanctions ended. Which would allow him to start work on billions in arms deals with Russia, oil contracts with Germany, France, China and Russia. You might note that these were also the nations most opposed to taking down Saddam.

But don't let objectivity and reality get in the way of that smug feeling you get when you bash Bush. i'll tell the girl at the campus bookstore you said all the talking points like a good sheep.

Re:We need real leaders, not computers

[Thomasje]

"A single event and the responses by inept leadership led to a variety of disasters that nothing and nobody could have predicted."

The event, 9/11, could easily have been predicted -- heck, it wasn't even the first attack on the WTC. Predicting the details of the second attack was impossible, but that there would be one was a given, given that militant extremism in the Middle East never showed any signs of decreasing and given that the U.S. never changed its policies in that region significantly.

As to the excessive, knee-jerk reactions to 9/11, including the systematic undermining of civil liberties, that should not have come as a surprise to anyone either. People *always* react like that -- when they get scared they'll agree to all kinds of extreme measures, and once the fear has subsided, it takes forever to undo the damage. People who remember the RAF terrorists in 1970's Germany can tell you how a bunch of armed college drop-outs, who were never anywhere near as crazy or violent as al-Qaeda, provided politicians with all the excuse they needed to erode civil liberties in that country just as badly as what Bush and his cronies have done in the U.S.

Re:We need real leaders, not computers

[DocSavage64109]

I like how you bring up what are now considered the "real reasons" of the Iraq war and then insult anyone who doubts Bush. Are you really saying that his and his cohorts outright lying about the entire purpose of the war and jumping on the 9/11 bandwagon are excusable, and that calling out those lies isn't objective or reality? Who exactly is the sheep here?

This has already been done

[mcsqueak]

42!

The unpredicable is predictable...

[Zeekamotay]

... at zombo.com.

did it predict the radiohead pirating scenario?

[wardk]

if not, then this is obviously no more useful than a pile of slag

does it run hot enough to keep pizza warm? there is a need for that

Stability

[mechsoph]

If the system is stable, you iterate until the error is negligible. If it's unstable, you spend your grant money on hookers, blow, and a ticket to Mexico, though perhaps not in that order.

Re:Stability

[Ristol]

If the system is stable, you iterate until the error is negligible. If it's unstable, you spend your grant money on hookers, blow, and a ticket to Mexico, though perhaps not in that order. Now that is what I call a Seldon Plan!

Ah well.

[Colin Smith]

You see.

People will trust the answer a bit of software gives them when they won't trust exactly the same answer, calculated in exactly the same way but presented by the same expert who wrote the software in the first place...

Particularly if the software system cost 8+ figures.

Isn't this not just...

[davidsyes]

unpossible? I can imagine the computer saying, "Insufficient data to arrive at a logical explanation."

Re:Isn't this not just...

[davidsyes]

I mean, can a computer predict how much toilet paper 25 different people will consume after eating at Taco Bell?

Butt, a head start would be to print on each sheet "Did I predict the number of sheets and plies you just used? Call 0BADF00D 000hDW1PE"

Re:Isn't this not just...

[Surt]

I'd be interested in knowing just how much toilet paper 25 people would be willing to consume as an alternative to eating at Taco Bell.

Maybes...

[Merovign]

My best guess is that, with enough data and broad enough terms, some kind of vaguely accurate predictive "crowd behavior" model could be made. Better than nothing, but not very specific.

There's already a lot of documentation on group/crowd behavior, and some particular people have been quite adept at "crowd control," which implies that there is a way to do it. Advertising firms sell their services based on the principle.

The A to B part of it with regards to computer software is the hard sell.

So... maybe, maybe not.

In a sense, this is just a sort of "expert computer" that mimics a human's skill at manipulation or control of crowds. In this context, prediction is intertwined with control - you need to predict to control, but once you control you can predict.

I doubt lack of data is a problem

[GuyMannDude]

Apparently there are those that have forgotten the old computer law of "Garbage In, Garbage Out". Even if we had a perfect model to predict these sort of things, we don't have any way of supplying the required data to model the prediction. What's the computer going to do, go undercover in secret groups? Read the web sites? Listen to radio chatter and analyze their conversations?

I think it's safe to say that a professor remembers something from Comp Sci 101. The article makes it difficult to know what they are truly doing, but their use of genetic algorithms and game theory indicates that they are hardly trying to build a "perfect" prediction model. I don't understand what your phrase "model the prediction" means, but it seems you are concerned about lack of data. The article specifically mentioned that there is way too much data for any individual or team of individuals to completely analyze. As far as how this data is obtained, the article doesn't go into too much detail but I think it's reasonable to assume that the inputs are going to be a combination of human intelligence (HUMINT -- "go undercover"), natural language processing ("read the web sites"), and communications inteligence (COMINT -- "listen to radio chatter", ala eschelon). Even if this cannot be done automatically, a human intelligence analyst can process the raw data and file a report with the system. The article doesn't sound like they are building lower-level analysis modules but something that will perform at a higher-level on exploited data.

One concern I have about the system is whether it will be effective against adversaries that rapidly change tactics. The insurgents in Iraq are quite clever in learning from their mistakes and are also quite innovative in trying new stuff. We may have a wealth of data on what they were trying last month, but it's not clear that this will help us predict their new tactic du jour.

Maybe someday when we have a real science of A.I. something like this might be possible, but all it shows is that this university professor will happily take government money for delivering absolutely nothing.

That's a pretty awful thing to say about someone based on a press release. It's reasonable to state you don't know how this would work. It's another thing to accuse someone of dishonesty. It truly saddens me that bold, baseless insults like this get modded up so quickly.

GMD

Re:I doubt lack of data is a problem

[Surt]

If you think it's safe to say that a professor remembers something from Comp Sci 101, man did you ever choose the right university.

Re:I doubt lack of data is a problem

[Reality Master 101]

I don't understand what your phrase "model the prediction" means, but it seems you are concerned about lack of data. The article specifically mentioned that there is way too much data for any individual or team of individuals to completely analyze.

To make a prediction about the real world, you have to model reality. Sure, there is a "lot of data", but that doesn't mean that 1) the data is useful, or 2) the useful data can be interpreted by a computer.

I think it's reasonable to assume that the inputs are going to be a combination of human intelligence (HUMINT -- "go undercover"),

Sure, but that's fuzzy data that can't be interpreted by computers using our current level of technology.

natural language processing ("read the web sites"),

There is no such science as natural language processing beyond grammar and syntax analysis. No models exist at all for doing useful-level interpretation of context and meaning. They already have phrase scanners.

and communications inteligence (COMINT -- "listen to radio chatter", ala eschelon).

See previous point.

Having a computer do tactical analysis, much less strategic analysis, is so far beyond our current technology it's ridiculous.

That's a pretty awful thing to say about someone based on a press release. It's reasonable to state you don't know how this would work. It's another thing to accuse someone of dishonesty. It truly saddens me that bold, baseless insults like this get modded up so quickly.

Granted, it's just a press release, but it's a pretty specific press release by the university itself (i.e., not filtered by the mainstream media). Maybe this is just trying to model troop movements (which you *could* do on a computer, given terrain, etc), but it specifically says, "social, political, cultural, military and media influences". Not to mention I get suspicious when they start hand waving phrases like "massively complex set of computer algorithms".

Again, maybe this isn't what it seems to be claiming, which is human-level pattern analysis. Maybe this is just a normal data-mining application that's been fluffed up by the university press release crew. But it smells like snake oil to me. Just because someone works at a university doesn't mean they aren't simply interested in feeding from the public trough. It happens all the time.

It'll have great military applications

[Centurix]

Major: Sir, the computer has given us a plausible scenario for operation Sandy Whirlwind
General: OK Major, lets see what this pile of junk has to say for itself
Major: It says that we can overcome all undesirable outcomes by sending in CL22 using a classic scissor movement.
General: Let me see that! How did the computer even know about CL22, our crack regiment of killer circus clowns! That's amazing!
Major: There's more sir. It also talks about project CC.
General: Project CC! The stealth car capable of carrying thousands of CL22 troops in a vehicle the same size as Robin Reliant? How did the computer even know about that project, it's only been discussed between myself and my 2 year old daughter!
Major: This program is amazing sir. Have another star.

sure here we go:

[snuf23]

Partially cloudy, scattered showers on the windward side, highs in the mid 80s. I live in Hawaii, 90% of the time weathermen who live hear have it easy.

predictions are easy....

[bigbigbison]

You can predict anything you want. You don't need a fancy computer to do that. Just flip a coin.

Making accurate predictions is something else all together.

Stock Market

[Anonymous Coward]

If you had such a machine, why tell anyone? Just predict the closing price for the S&P 500 each day and become a bazillionaire.

Observer Bias

[animusCollards]

- "Mr. President, our computer model says you won't invade Iraq."

- "It does, does it? I'm from Texas, boy; no Gosh durn computer's gonna tell me what to do! We start bombing in five minutes! (Dang, I always wanted to say that.)"

If the program's output is available to one of the "players" in the scenario, how does the algorithm account for that?

...or do you simultaneously upload the program to WOPR & Nomad? Hmm...

The Movie "Paycheck".....

[3seas]

...good movie with Ben Affleck and Uma Thurman.

Tell the future and then make it happen...

But then there is the quantum physics problem of changing the outcome by observing it.

At what point do you prove the software actually works in a manner that doesn't lend to the creation or alteration of what would have been had it not been predicted in the first place?

Oh I know, 2.2 million to produce software to predict the future but nobody is allowed to see the results.

Or this is where computer become smarter than humans by making humans not need to think for themselves...Hmmm, some already don't and they apparently have loads of money...

Re:The Movie "Paycheck".....

[mrbobjoe]

...or the short story "Paycheck", by (as with the Minority Report references further up) Phillip K. Dick. I thought the movie was one of the better story to film translations of Dick's work, though, so not too much in the way of complains about how they're completely different.

The art of proposal writing

[thethibs]

I thought that I was a pretty good proposal writer, but this guy has me nearly speechless with awe! I bow to the Master.

They should hire Think the Unthinkable...

[Pearson]

It would make as much sense.

Think the Unthinkable

Maybe it uses Markov chains

[Merusdraconis]

Can it predict this response?

Okay, I guess it is a predictable response, but can it predict this banana pants?

In 7 months...

[Anonymous Coward]

Just wait until this massive agglomerator of news encounters April Fool's.

Or next week's issue of The Onion.

Comment removed

[account_deleted]

Comment removed based on user account deletion

I predict...

[Anonymous Coward]

Failure: Event Code 00001

I'm guessing...

[qzulla]

They never heard of the chaos theory.

  Deep Blue, the first computer program to beat a world chess champion, is an example of how ATRAP can respond to changing factors, Ten Eyck explained. "Every time its opponent made a move, Deep Blue recalculated all the possibilities and likely courses of action, eventually settling on the fittest move that would achieve its goal of winning the game."

However, chess is not an exact analogy because only two players are involved and the end goal is for one player to win.

In unstable areas, winning often means establishing an environment in which the factions co-exist in a win-win situation or at least in an equilibrium in which there are no rewards, and some penalties, for disturbing the status quo, Rozenblit said.

"Deep Blue is a good analogy because it illustrates the complexity of the problems, but in chess you have a finite court and a well-defined set of operations," Rozenblit added. "Therefore, a move constitutes a valid move.

But what we're dealing with now is a world with no rules, with infinite possibilities and moves that defy logic, such as total disregard for the basic instinct of self preservation."

Or maybe they have but are ignoring the fact it cannot be predicted. I like the last graf. It kind of says it all.

Oh well, good luck on that one.

qz

It could happen... or not.

[Anonymous Coward]

Meh, gimme a shout when it can dream the impossible dream.

The driving term of the equation

[zazenation]

It's intuitively obvious that any mathematical model which attempts to predict future events by analyzing past data sets would have to be driven significantly by the percentage of women on the planet whose menstrual cycles were coincident on the day or days in question. Just imagine the ripple effect on those unfortunate enough to possess a Y chromosome. Maybe there could be some wavelike effect or menstrual probability distributions that follow chaotic menstrual Attractors, which when aligned, bring catastrophe and misery during "perfect storms" of PMS. That actually fits --- Perfect Menstrual Storms . . .

Or maybe these guys have found the brass ring to the money tree for all eternity.

My 2 cents with a wink.

Version 3.0

[Nazlfrag]

10 for i=1 to 4
20 read a$
30 print a$
40 next i
50 end
60 data "You're not going to like it", "You're really not going to like it", 42, "I think the problem is you don't know the question."

Not weather

[volpe]

I suspect that what this software can predict is not so much like the weather as it is like the climate.

It's a what?

[glwtta]

The Asymmetric Threat Response and Analysis Project, known as ATRAP

What is this, Fark?

Prediction

[Anonymous Coward]

Great for predicting macro changes within a large population.

Seems worthless for anything else.

What happens when an advasary steals your predictor or develops their own to gain insights into your reaction in a way that effectivly counters them?

When does it start printing out the message "There is another system"?

In further news...

[jd]

...it was revealed that GALLUP and MORI are to beta-test the new product when it comes out, as it can't be any less accurate than surveying one ten-thousandth of the population on dubiously-phrased questions. Meanwhile, speculation that ATRAP will replace the current ballistic missile shield by printing out sufficient quantities of meaningless data to trap the missiles in the world's largest paper jam have been denied by Admiral Poindexter.

....wells

[wellingj]

The software will predict the actions of paramilitary groups, ethnic factions, terrorists and criminal groups, while aiding commanders in devising strategies for stabilizing areas before, during and after conflicts.

I've got a simpler solution that can be implemented today. As a US citizen (which I am) I can realize that it's none of my goddamn business what other 'areas' do. OTOH, Katrina sucked. But I doubt any policy maker will listen to such software, because they barely listen to their constituents anymore.

Who cares?

[JK_the_Slacker]

I don't understand what impact such a program could have on the foreseeable future...

Lotto anyone?

[Anonymous Coward]

If they are smart enough to invent it, they would have been smart enough to use it on lotto result and make way more money.

Depends on the situation

[mveloso]

Your analysis seems valid, for situations that are relatively stable. For system which are in flux (such as in a combat area), reality is substantially more fluid than, say, the traffic patterns in Queens.

It takes events like 9/11, or the invasion of Kuwait by Iraq, to adjust the normal state of affairs. In a flux situation, small actions (and individual actors) can cause tremendous instability...or crystalization, depending.

Re:Depends on the situation

[KDR_11k]

Hell, imagine someone predicting the year 2000 from the year 1900. Even if they can accurately model all the societal responses they cannot predict technological progress. If the atom bomb had never been invented the cold war would not have stayed cold, a prediction would not have known about the atom bomb beforehand and would have predicted that the war would break out.

Re:Depends on the situation

[WeeLad]

Would the computer have expected the Spanish Inquisition?

Re:Depends on the situation

[bmorton]

NOBODY expects the Spanish Inquisition!

First pronouncement

[Anonymous Coward]

The program's first announcement will be "This will never work."

Followed shortly thereafter by it bursting into a paradoxical ball of flames.

skynet?

[jakepmatthews]

didn't i see this in the terminator? i didn't work out so good for them...

Leader

[duggi]

Why bother with a computer in the first place? Are we so dumb that we cannot predict the reactions to some happening? Look at any good leader, and he can predict how exactly are people going to react to his actions. All dictators did that. This is something which takes experience, wisdom along with knowledge. Isn't it better that we take the easier route? Why are we hell bent on making computers do all the stuff? They just compute, let them do it, why bother it with stuff we ourself can handle?

Mod parent up.

[TheLink]

That's what it's for.

Predictable ;).

just pondering

[Polir]

There was an article about quantum theory. And it said that there is a multiverse, and at each questions/answers the universe splits into two for one and the other version. Also it was saying that when you measure anything you change the subject by the measurement. Before it is measured there is an uncertain state and the measurement will make it certain. If the above theory which was said to be commonly accepted by the big names of the field, does that mean that if you build such a computer which lets say works 80% perfect, that by computing the answers it will also influence our reality? I mean not by obvious political abuse but in sense of quantum physic/scientific means?

Isn't that what Vista's supposed to do?

[buttle2000]

Am I going to use my pc? rip a dvd? share a file? change hardware? plug in a dv recorder? play media? copy?

I'd say I'm about as unpredictable as any old politician/terrorist who want's to: have more power? control people? invade? subvert? violate?

Some days ago here on /. I read that M$ thinks it's antivirus tecnology can be used to fight against AIDS.

Is seems that, once again, MS tecnology is really out there on the forefront and Ballmer just missed out on being awarded $2.2Million to develop software to predict the unpredictable.

That's just great

[G-News.ch]

Now with software deciding what is right and what is wrong to do, the US can have presidents even more stupid than George W. I can also imagine how the software will manage making a decision based on millions of factors. You just weigh factors accordingly. If an individual life for example has a factor of 1, a barrel of raw oil would have a factor of 10. Then equations where you'd have to decide between keeping alive a million civilians and getting 10.5million barrels of oil suddenly seem very easy to resolve. Seriously, once computers start making policy, how far from "SkyNet" are we then? Why doesn't it surprise me that this program was launched in the US? If you think a computer can decide better what to do than your politicians, it's about bloody time to replace them.

Confusing Herding With Specific

[SerpentMage]

You are confusing two aspects. General, and specific.

Here is the general.

>Equivalent arguments apply to the "free will" of individual human beings. Humans tend to congregate in packs - behaviorally, philosophically or otherwise. This strong tribal leaning that is presumably built into our genes ensures that most behavior patterns will be statistical in nature. Indeed, the actions of an individual can be simply predicted to a first approximation by merely qualitative means even in the absence of complete information by assuming rational behavior. A better approximation can be achieved by modeling the level of rationality of the individual and assigning probabilities based on that.

This is like saying tomorrow 35% of people will eat eggs for breakfast. This is human nature at its best. This is also what sales companies follow.

What the 2.2 million is aimed at is figuring out of if tomorrow people will start grazing like cows on the field. This is called a rare event and something that many people would like to figure out. In fact this is what sales companies would like to figure out and they can't since it is RANDOM for lack of a better word.

You can't predict the unpredicted because that is like saying I can predict tomorrow's lotto numbers. Here is the problem, and this is the problem of the program. There are literally several billion combinations, and the computer software will tell you the same thing, in that there are literally billions of things that people can do to inflict damage.

So you have the answer, and now comes the question what do you do? Do you follow all possible leads? Do you implement all of the solutions? The problem is that you will get answers like 10% this, 5% that, and 25% over there. Each of these answers is completely valid, yet which do you follow? Do you follow 25%? Yet then the 5% is possible, and while 25% is greater than 5%, 5% is still possible.

What people fail to grasp is random. Let me illustrate. There is a statistic that certain storms are 1 in a 100 years. So when can this storm happen? It can happen next year because you have 1 and a 100 years. Then when the 1 in a 100 years storm happens, when can it happen again? Again next year, thus in two years you have 1 in a 100 year storms because that is the way that random happens. Though after that you might not have a storm for another 200 years. Or if you do then the statistics have to change. But notice what happens? Statistics are after the fact, not future predictions.

If you want to see this action for yourself flip a coin. Do you get HTHTHTHT? No you get HHHHHHTHHHHHTTTHHHH, and if you flip often enough you will get a 50 50 distribution. So now comes my challenge predict what you will get, H or T. If it is fair and a random flip you will probably be 50% right!

So what about the 50% wrong? This is a long term statistic, but at the singular level where life and death is involved your odds are 50 50.

This is a fools game and indicative of what quants are trying to do...

aid commanders?

[hitmark]

dont they know that these kinds of things works best when its wired up to do the commanding directly?

Bhah, Use the force Yoda

[Anonymous Coward]

At the very best they might get a very fuzzy picture of what *could* happen, but people are not very predictable.

Predicting self...

[Anonymous Coward]

What would happen if you fed the program to itself?

Re:Rozenblit's work on coevolution/genetic algorit

[Anonymous Coward]

The abstracts alone can't proof anything. And I personally think that genetic algorithms suck because lack of mathematical foundations.

What's predictable

[tchi.keufte]

What's predictable is that no computer program will ever be able to cook the uncookable, or predict the unpredictable, as you like...

for this kind of stuff

[kaizokuace]

I think some sort of algorithm (computerized or not) that manages a system of human analysts, strategists, experts in relevant fields and the like would work out better. The key to taking in real world data and analyzing it involves taking trends and other vague generalizations and piecing together a sort of quilt that gives a better view of what is happening, has happened and as such what what might or will happen. Humans are fairly simple when it comes to actions they take. Sometimes it is unpredictable but for the most part people do similar things in similar situations and for the most part these actions are based on the wanted and unwanted in life or in a situation which stems from the general human experience which is at its root the same for all people (birth, growth, death, resource scarcity and a need for those resources such as food, protection, social contact and many other things).

PROFIT!

[hanshotfirst]

This will be HUGE in corporate IT departments. IT Managers always want to know when the next unplanned outage will be, or to identify/correct the bugs which have not been discovered yet.

Donald Rumsfeld

[Calathea]

This will be handy to find the known unknowns that we dont know of that Donald spoke of.

This would could come in handy...

[JaumPaw]

...As a nice addition to the hitch hiker's guide.

Expect the unexpected.

How will they tell if it works?

[chord.wav]

How will they tell whether the system is working correctly or not? What if it predicts a ton of outcomes that never happen?
Will it be considered reliable if the predictions happen?

So, it (or its constructors) could make sure the predictions happen. If they don't, it'd seem like a waste of money. So, it predicts a flash mob, and at the same time, it starts it. Like a baby Skynet.

easy

[AlgorithMan]

> Predict the Unpredictable

cat /dev/urandom

do I get $2.2Million now?

Music

[Stooshie]

A nice analogy when people think computers can make decisions or have "Artificial Intelligence" of any merit.

Pupil (Excited about AI):- I have just written a programme that writes music in the style of JS Bach.
Tutor (Seen it all before):- Really? How does that work then?
Pupil:- I programmed all of the known manuscripts by Bach and the computer uses that to write new compositions.
Tutor:- Great, can it write in the style of Mozart?
Pupil:- Sure, give me all the compositions by Mozart and I'll show you.
Tutor:- You mis-understand, can it decide, of it's own volition, to write in the style of Mozart.
Pupil:- Well, no it needs to base it's composition on something.
Tutor:- It has the entire works of Bach, is that not enough?
Pupil:- No, it needs the entire works of Mozart to write in the style of Mozart. Hell, even music students need to have heard Mozart in order to write in the style of Mozart.
Tutor:- Oh, so how did Mozart do it then?

Re:Music

[Anonymous Coward]

Sorry but that's a poor analogy.

To create, one does not need previous examples, to imitate, one does. This is not an indication of intelligence or lack thereof. For all his genius, Mozart would have been no better at imitating the software's "style" without having heard it.

Re:Music

[Stooshie]

... To create, one does not need previous examples ...

Erm, that was my point. Mozart didn't need previous examples to write like Mozart. The computer does. The computer can only imitate and cannot, of it's own volition, decide to write like Mozart (having only heard Bach before).

Of course Mozart had heard Handel and Vivaldi and Corelli etc. before, as well as experiencing love, hatered, jealousy, dissapointment etc. But, he still wrote like Mozart without any prompting.

Re:Music

[Stooshie]

Just seen your last point.

Any computer will only imitate. It cannot decide to do anything of it's own volition. Mozart can.

Even if you fed Handel and Correlli and Vivaldi and Hundereds of other composers into the computer it would only write a mish mash of the composers fed in, following a set of rules in it's code.

Mozart would add extra to the music. In computer terms he would create new rules without any prompting, without even being told he could create new rules (or break them).

Re:Music

[Anonymous Coward]

The analogy implies that Mozart is more "intelligent" because he was able to create without previous input. My original point was that creation does not require previous input while imitation does. These can be viewed as different activities and it does not follow that the computer which is less intelligent because it is not creating. The stated goal is different. If Mozart is given the same goal (imitating) he would perform no better than the computer without some kind of exposure to whatever he is trying to imitate. He could be making the most beautiful music in the world, with all the rich influences you cite, but it would be unlikely to resemble some specific style unless he had a "base for composition" just like the computer or student musicians.

In regards to your subsequent posts, I agree that humans currently remain better at creating for humans than machines. But this is not what the analogy is demonstrating. Especially because, in the analogy, human students likewise imitate. This implies that the ability to create stems not from being human, but from being Mozart. This does not seem to be the point you were trying to make, which is why I disagreed with the analogy.

As you mention, Mozart did not create in a vacuum. You could argue that his "programming" included not only the entire musical repertoire to which he had exposure but also the biological, social, and cultural circumstances in which he lived. In that sense, creating a Mozart is as tedious for our species (evolution, society, and so on) as feeding data into a machine. What he created could just as easily be described as a mish mash of his exposure and experiences, based on the cumulative effort of the whole civilization before him, with the only difference being that society and nature had was feeding him data instead of a programmer. Current machines require a lot of hand-holding of course, but it is not obvious that they always will. Attach a microphone, give it some way to get around, and suddenly you will start hearing influences of birdsong, traffic noise and whatever's on the radio in its music. What a machine will probably continue to struggle with is the ability to evoke emotion in their human audience. Mozart can listen to his own music and understand that a certain passage evokes dread. The computer will have to rely on the audience to communicate that result to it, assuming that it is unable to "feel dread" itself. I don't know if this is an unsolvable problem. Maybe it's not a problem at all if the range of responses from an audience is sufficiently diverse. It might be argued that just because an artist believes that they are evoking a certain response in their audience doesn't in any way correlate to what the audience really feels. If so, the machine would be the ultimate haughty artist. It makes what it wants and the audience can take it or leave it.

At that point, all that is necessary is to make a distinction between creating, and creating something of value. This can be kind of arbitrary. If all that is required to create is that it be novel, spontaneous and unbound by rules, the addition of a random number generator can make a cutting edge surrealist out of a very simple piece of software. On the other hand, if you add the requirement that creations also be understandable (i.e. based on established rules of language, composition, experience, etc.) and valuable to humans, you require some kind of fuzzy logic which can establish a range of acceptable output (e.g. all tones must be in range of human hearing) but can be flexible within those constraints. This is a challenge of course, but not impossible. It is also not so different from a human creative process in my opinion. Pop music operates very much along those lines, as do formulaic sit-coms or novels. Draw a frame, then fill it in with some established techniques. Their value artistically may be debateable, but socially and financially they are obviously of worth. Can the same rules used to make pop tunes be generated or modifie

Darren Brown

[oliverthered]

although most of his stuff is probably trickery anyone whos watched Darren Brown should know that humans are fairly predictable.

wait...

[Anonymous Coward]

till i use it to interface with Infinite Improbability Drive....

That's so Asimov

[Chris whatever]

Yeah yeah Psicho-history and the study of the mass population to predict the future.

Foundation's Edge

[kcdoodle]

Already been done in science fiction. But just add one Mule, and the results get severely skewed.

Yes, but...

[drunken_boxer777]

Can it predict the Spanish Inquisition?!

Can it predict..

[Anonymous Coward]

... what day of the week it will be in seven days?

But will it play poker?

[ubergeek65536]

Sounds like it may be good at calling a bluff. A collegue and I were recently talking about how a good poker engine would need to be able to predict irrational behaviour.

This assumes that anti-US actions are "irrational"

[Der Einzige]

Two money quotes from TFA:

The software will predict the actions of paramilitary groups, ethnic factions, terrorists and criminal groups ...

Since the end of the Cold War, our opponents have behaved in ways that defy what we would consider normal logic, pursuing actions that we find almost inconceivable ...

Note especially the text I bolded. This assumes that "normal logic" predicts actions that conform to American cultural and ideological biases.

This is not about predicting truly unpredictable behavior. This is about predicting behavior that is perfectly logical from the viewpoint of the people doing it. US planners can't admit that the behavior is logical, because that would be granting rational legitimacy to opposing US government policy, and that would mean admitting that the US is not always right.

This project would never have been thought of if we tried to understand our opponents instead of demonizing them. This project is only necessary because the US is committed to pretending that only irrational people oppose US policies.

If, for example, Americans understood how Muslims feel about the Holy Cities of Mecca and Medina being defended by infidel soldiers, they wouldn't be so quick to call bin Ladin and al Qaeda irrational.

US writers like Thomas Friedman and Victor Davis Hansen try to diagnose militant Islam as "irrationality" and "immaturity," because they refuse to consider that anyone might have rational reasons for opposing US military and commercial power, or for trying to defend their traditional culture and religion from US cultural influence.

Note that I am not saying that America's enemies are "right." This is not about right or wrong; this is about competing political and cultural interests.

Whatever money is going into this project would be better spent on hiring more State Department people who speak foreign languages and understand foreign cultures.

Put down the Sci Fi and back away slowly

[snowwrestler]

Asimov was a fiction writer who, in real life, was no more adept than others at predicting the immediate future. His accuracy in the far future remains to be seen.

Despite your rather vague analytical analysis, the fact remains that the details of even subsets of human endeavour remain strikingly unpredictable over a wide range of scales.

Can we predict things in a highly imprecise way? Yes, probably. I could say that the United States would still be an operating and autonomous nation-state 10 years from now, and I would probably not be wrong. I could even predict that the President 10 years from now would be either a Democrat or a Republican, and would probably not be wrong.

But are these predictions actionable in any real way? Not really. In the same way I can predict that it will be raining somewhere on November 2, 2008 (U.S. election day). Almost certainly true, but of course it doesn't really answer the question of who, specifically, needs to bring an umbrella to the polls that day.

In politics and finance there are ample opportunities for those who can prove success in detailed, accurate predictions. A man who can predict the future of even an individual stock one day in advance, reliably, can become very rich. Yet, there are very few who actually succeed at that sort of thing, especially over a long term.

Rudimentary control system for...

[tiqui]

the Infinite Improbability Drive.

Questionable definitions

[paladinwannabe2]

Because no government in history has ever significantly and permanently reduced its power or revenue through the process of democracy, and certainly not through the process of bureaucracy.
What about the Chinese government? It reduced its horrific abuse of human rights to a mere wanton abuse, all through the power of bureaucracy. I'm also assuming you don't count local governments, in which you can find hundreds of examples in U.S. history of the crime lords being voted out of office through democracy. Not to mention all the times when everyone in a government has been voted out, to be replaced with less obnoxious politicians. There's also things like the amendment to end prohibition, which is about as close as you can get to a permanent reduction of power in a democracy. Things like the draft are unlikely to be used again without another World War. CONINTELPRO was illegalized, as were the Alien and Sedition acts. 1954 had income taxes as high as 91% in some tax brackets... does lowering that count as permanently decreasing power and revenue?

The statement sounds pretty, but when you're dealing with something like democratic governments (which have a tendency to change radically over the years) it's also pointless. You might as well say 'No Roller Coaster has ever significantly and permanently changed its distance from the ground'. There's no way to EVER say a change is 'permanent' in a democracy, because the next people in office might further reduce (or increase) the power of government. There's even debates about what constitutes 'power'- does a law forbidding slavery increase the power of the state or decrease it? Do laws limiting tariffs increase the power of the government or decrease its ability to earn revenue? How about laws that enforce equal treatment between races?

In short, your generalities are baseless, pointless, and probably incorrect besides. Power does tend to naturally flow to the hands of the who already have it, but democracies are the best way to ensure it doesn't stay there. Of course there will be ups and downs, but if you look at U.S. history, you'll see that everyone other than a white landowning male has more power relative to the government than they did when this country was founded. Furthermore, even the white, landowning males have more power relative to the government than they did during World War II, and probably during most of the time since. Things are looking worse for this country than they did 10 years ago, but I'd rather live in this age of government than the government 50 years ago. (Not to mention the government 50 years before that).

Don't tell the Poli Sci department...

[Loosifur]

My degree's in political science, and I concentrated in international relations theory. This just sounds like applied game theory with all the switches thrown. One of the problems is that, while you can account for players lacking information in theory you can't account for what information the players lack in a practical, real-world application. In other words, a player might make a subjectively rational, objectively irrational decision based on imperfect information, and in a controlled environment we'd know what the imperfect info was. However, as one of the players in the game in the real world we wouldn't necessarily know what information we lacked. It's easy to predict a poker game's outcome if the table's glass and you're underneath it. Not so easy if you're one of the players.

Another major problem is that, at least when it comes to IR, there's still a lot of debate over who even makes the decisions. Some theories say that only states matter, and that they act monolithically, some say it has to do with historical trends, some say it's inherent in the capabilities of states (i.e. powerful states become dominant, as opposed to domineering states becoming powerful), some say it has to do with culture, economics, individual personalities within political systems, the structure of the state itself, etc.

It's an interesting idea, but I think it's doomed to fail considering we don't even know how to phrase the question yet.

it worked for the world cup of football... NOT!

[C0vardeAn0nim0]

http://archive.gulfnews.com/articles/06/06/07/10045272.html

two dubai students created a computer model that said brasil would beat italy on the finals to win it's sixth title.

unfortunately, the model was unable to predict the humiliating defeat we sufered to the french (we played francec three times on world cups. lost all three, including a final) and italy's victory on the penalty shoot-out over france.

if the model correctly placed italy on the finals, that was more a lucky guess than a correct prediction.

football (not the america kind, mind you. the regular kind) is by definition unpredictable.

Never possible

[SCPaPaJoe]

If you can predict the unpredictable then, by definition, it isn't unpredictable.

Right?

Who needs this to define foreign policy?

[oneopinion]

Hasn't the president already demonstrated a potent ability to "predict the unpredictable" and generate outcomes that defy logic? Who needs a computer algorithm to do it?

Free Will - no such thing

[PMBjornerud]

People who tout their own "free will" should think long and hard about that and realize that simply being able to imagine a multitude of choices does not mean that each is likely to occur. Well... Free will is impossible.

Where should such free will stem from?

1: Quantum randomness? Unpredictable, maybe. Free? Nope.

2: Your experiences and cognitive choices? Shaped by genes and environment. Missing a bootstrap, there is no point when your free will can make a descion that later choices can be based on.

3: A soul? No. It could be:
3a: two-way linked, meaning your genes and environment could impact your soul, thus change what was supposed to be your free will.
3b: one-way linked, meaning your soul cannot be impacted by your life in any way. This means the soul cannot care about what happens to you, which renders the whole idea utterly meaningless.

Obviously, I'm not going to disprove the concept of free will in a 6-line argument (plus linebreaks). Just posting a summary if anyone was interested in pondering it. I'll try to refine it into a zen-one-liner over the next 10-30 years. anyhow, it's pretty clear that "free will" is an utterly impossible concept.

Which is fine. Then you can realize that you are a marvelous, complex being, capable of organizing and abstracting life from the swirling vortex of unfathomable mayhem and chaos that the universe is on the nano scale.

We've got something way better than free will.

Re:computer? spoiler alert

[cellocgw]

What everybody forgets is that Asimov undercut his own psycho history by introducing a random event, the Mule.
And further, what few ever knew is that one of Asimov's last books ..

SPOILER ALERT

it turns out that a lot of the 2nd Foundation work, and general evolution of the Galactic Empire was invisibly organized by Daneel Olivaw. So much for mass chaotic destiny.

bet no computer would predict this

[Anonymous Coward]

http://goat.cx/

go ahead and click the link....it's not what you think....really!

Re:Rozenblit's work on coevolution/genetic algorit

[FleaPlus]

What do proofs have to do with it? They're engineering papers.

no

[slew]

Not true.

Giving kids everything they could possible need makes them _spoiled_
Giving kids everything they could possibly want makes them _entitled_

Eventually, a kid will need to get things they need by themselves. Delaying a kid's recognition of this fact will make them spoiled (at least a little bit in the best case). In addition, giving everything they could need will deprive the kid of the ambition and self confidence they would gain from doing the things they need to get done by themselves. Of course as a parent, it's purdent to provide a safety net in case things don't work out as expected and it doesn't hurt to give them _some_ things that they need or want, but there needs to be something that kid needs to do to grow up and be a contributing member of society (and it seems to me that that's a parent's primary job).

Of course another way to approach this is to want your kid to be dependent on you for all their needs the rest of their lives (I know parents that desire this type of outcome, so it's not actually a rhetorical statement).

Feel free to substitute parent-kid with government-citizen at your convenience...

probably not

[slew]

Let's say I roll a pair dice and "predict" that it will not add up to 6 (about ~80% perfect).
Let's say I took in to consideration many measurements on the pair of dice and the table and the local gravity fluctuations, but I always predicted the same thing.

Despite what your craps playing quantum-physicists may think, I don't think there's any quantum theory that any meaurements not made on quantum-timescales could affect the roll of the dice or alter reality.

However, if in the highly unlikely coincidence that this magic prediction device will be a quantum computer which discriminates the quantum-states of macroscopic events before quantum decoherence kicks in, perhaps there's some small chance that it will influence reality for a short time. But the current thinking is that quantum-decoherence happens on really, really short timescales for macroscopicly observable systems, so it probably wouldn't have more effect on reality than if you flipped a coin to decide whether you go to the movies or to a concert on Saturday (or relied on the magic 8-ball).

Let me know...

[bratwiz]

Let me know when you get that working, I'd like to buy one to start my new brokerage firm with.

Oh, and I'll pay you with the profits I'll make using the machine. You designed it, you know I'm good for it, right???

Oblig. Admiral Ackbar

[joeslugg]

"It's ATRAP!"