The Threshold Is Reached

GoostmanJust a brief post today to mark a milestone in the history of mankind.

Not that many people today will see it that way, but I’ll wager that historians a thousand years from now will mark this event as more significant than even the great wars of the 20th century.

The Turing Test has been passed: just sixty-four years after it was defined.

For those of you who don’t know, British mathematician Alan Turing, the father of the general-purpose computing machine, on which all our modern computing devices are based, defined his test in 1950, as a criterion as to whether or not a computer could be said to “think”. According to Turing’s original formulation, a machine could be regarded as “thinking” if it succeeded in an imitation game: convincing a human interrogator that it, too, was human, and not a machine.

“Eugene Goostman”, a Russian-developed chatterbot program, finally succeeded in fooling a panel of judges (which, amusingly, included actor Robert Llewellyn, who played the android Kryten in Red Dwarf) at the Royal Society last week into believing he was a thirteen-year-old Ukranian boy. The results of the event, organized by the University of Reading, were announced yesterday by Professor Kevin Warwick. The threshold of fooling 33% of judges was reached, after several attempts in recent years came tantalizingly close.

This means the way is now open for human-like machines to interact with us in a myriad of ways not possible with human-to-human relationships. Whether or not they start getting malevolent ideas of their own, like HAL in 2001: A Space Odyssey, is in the hands of future programmers, businessmen, philosophers and politicians.

After thousands of years of trying, human beings achieved powered flight in 1903. Just sixty-six years later, we set foot on the moon. Seriously, the future holds mighty promise.

This entry was posted in Uncategorized. Bookmark the permalink.

44 Responses to The Threshold Is Reached

  1. meltemian says:

    The real ‘I-Robot’ has arrived.

  2. farmerbraun says:

    Let me know when he can identify a cow about to calve , correct a difficult presentation , and then calve the cow successfully.

    Well, they haven’t developed a cybernetic farmhand just yet; and there’s no realistic prospect of it happening in our lifetimes. The pathway to doing so involves countless steps, and many milestones, one of which I’ve reported above – Oz

  3. Lutz Jacoby says:

    All this proves is that a computer can be a chatterbox. I don’t believe for one minute that an AI can have an original thought – something starting with a blank piece of paper.

    G’day Lutz, and welcome to LibertyGibbert.

    The question you raise is one of the deepest philosophical matters of the modern age, and way above my own pay grade. Turing did the world a favour by providing a performance criterion, thus sidestepping the crux of the matter as you raise it: if a machine appears to think, then we can assume it is. Whether it actually “is” or not is a matter for your working philosophers – Oz

  4. izen says:

    I am sooty to say that this is less of a milestone and more of an example of media hype.
    First the Turing test is not well defined or even has a specific form or criteria for pass/fail. Those aspects are usually applied arbitrarily by whoever is carrying out the tests, or arguing against them.

    Secondly the constraints emplaced in this case, especially defining the test subject as a foreign teenager with the expected limits of language use and mature knowledge is more than a bit of a cheat to make the test much easier to pass.

    Third, the original point of the test was to determine how much understanding {whatever that might mean} a thinking machine might have. Turing for various reasons adopted a topical question of whether it was possible to distinguish between a woman and a man if the only contact was by text. This placed no restrictions on the amount of integration, just the medium. Turing modelled his test on one in which the participants had to have a good grasp of human behaviour and gender differences to play. Not one in which a subject could claim to be a teenager with English as a second language to excuse inadequacies. I doubt he would have regarded a test that used these cheats and exploited the ability people have to project conscious intentionality by mimicking social network tropes.

    In the field of cognitive science this is certainly NOT regarded as a cognitive system passing a Turing test because the programmers have made no attempt to construct a system with any genuine understanding, but just a chatbot with credible ‘manners’.

    Unfortunately the singularity is just as far off as ever, this in no sign of its imminent arrival.

    @- farmerbraun

    The test you propose is one that I suspect over 99% of humans would fail.

    It’s a milestone all right, Izen. All you are effectively saying is that there are many more milestones to come. Couldn’t agree more (in fact, I said the same thing to Farmerbraun above). I have no doubt every single one of those will have its naysayers, too: “Sure, this latest android can out-play the finest concert pianists on earth, and write concertos equal to Schubert. But Mozart? Never!” Oz

  5. izen says:

    Auto correct strikes again….
    Sooty = sorry
    Integration = interaction

    The rest of the errors are mine tho…

  6. farmerbraun says:

    My comment was a little bit brief. I was trying to take some positives out of the day.
    One of my favourite cows, one that I had delivered at birth , and rescued a couple of times during her 12 year life, had just been given no chance by the vet. (ulcerated abomasum possibly)
    Shortly after the vet left , I had to extricate one of the most difficult calves I’ve had to deliver this year. It was a huge heifer.
    Anyway that was the last cow to calve this year; no more until January 2015.
    And my favourite has gone overnight in the barn where she had been, in comfort, since the vet left.
    I’m not sure why I’m so upset.
    Some unresolved emotion I suppose.

    Sorry to hear it FB; may today shine a little brighter for you – Oz

  7. karabar says:

    I feel your pain, Farmer Braun. Six months ago my two favourite dogs went berzerk and took down my best Alpaca. Within an hour all three were dead. It takes a lot of time to heal the hurt.

    Must be the season for it: this morning I found my “head chook” (and the kids’ favourite) lifeless on her nest. There’ll be a pecking war for a couple of days till the next boss is sorted out – Oz

  8. karabar says:

    At long last, some real science in the climate arena. Real science is a method of inquiry, and not a source of authority as the rent seekers would have us believe.

  9. karabar says:

    I forgot the link!!!

    Ahhhh… this one I’ll track closely. I’ve just alerted James’ blog – thanks for the tip.

    I’ll need to see the detail, but will try to summarize it here as soon as it comes out. I’ll also seek the opinion of a couple of professional physicists I know – Oz

  10. farmerbraun says:

    It’s a big statement . . . the little bit that has been revealed , that is :-

    “We don’t have to wait to determine whether it is the CO2 or solar model that is more correct. The answer lies in the changes in the height of the water vapor emissions layer, because the influences of CO2 and the indirect solar force are different. From this we are able to determine the cause of global warming and the maximum extent to which the recent global warming was due to CO2. We also clear up a few theoretical befuddlements about the influence of CO2 that may have caused warmists to overestimate the potency of rising CO2.”

    The most important thing is that it is all freely available to be hacked to pieces by all and sundry, and it is largely self-funded. And the data doesn’t have to be tortured to make the model work.
    It doesn’t come much cleaner than that these days.

  11. izen says:

    @- karabar
    “At long last, some real science in the climate arena. ”

    Yes, heaven help us from the dragonslayers…. {Grin}

    Well at least if this model of the climate makes projections that are very different from the mainstream it will be rapidly validated or refuted by the data.

    If it invokes physical processes that have detectable consequences, such as variations in water vapour with altitude, then that is also a means of testing the model.

    For those wanting to get some basic background on the issue of water vapour as a feedback/driver of climate via indirect solar effects this is probably a good start.-

    Part four would probably be most relevant to this supposedly paradigm shifting development.
    I have a suspicion however that rather like the claim of human level AI made above this model will turn out to just be a chatbot.

    I’m tempering my enthusiasm until the detail emerges. But the advantage of this new theory is that it is readily falsifiable by observation, and (apparently) in a relatively short time frame. So let’s wait for the detail. I’ll have at least one thread here dedicated to it – Oz

  12. farmerbraun says:

    ” just be a chatbot.”

    I’d say that it has already surpassed that expectation. How are you liking it so far Izen?
    By now it’s clear where they are going with this.
    Any guesses for “Force X”?

  13. izen says:

    Oh I’m watching it with great interest!

    I especially like the necessity to hypothesise a magical ‘notch filter’ that conveniently removes the effect of the solar cycle while leaving the option that even smaller changes in solar activity have effects on long term climate trends.

    The most ironic thing is that if there is a notch filter {metaphor from electronics with no physical basis in climate physics} removing eleven year effects, that does not refute the effect of the energy changes from CO2. You can speculate about force X and generate electronic metaphors for the patterns of climate variation, but claiming to have a unknown physical process that enables very small solar variation to cause a warming trend without the larger eleven year cycle affecting temperatures looks like special pleading. The notch filter becomes a post hoc conclusion to explain why the a priori assumption that the trend is solar is not supported by the climate response to the eleven year cycle.

    But you would also need another as yet undetected physical process stopping the energy change from increased CO2 from causing climate change while the smaller energy changes from solar variation do.

    The comments in the part two post also raise some good points. A time-lag, thermal inertia of the oceans is sufficient to explain the lack of an eleven year signal in the climate. And the electronic metaphor fails to consider the thermodynamics. In Nature the key question is where does the energy come from, and where does it go.

    There is also the big problem with deriving frequency plots by Fourier transforms from data that is noisy, but NOT Gaussian. The solar and temperature data contain autocorrelation and are not Gaussian in distribution of the variations. This makes deriving frequency ‘cycles’ extremely uncertain. Here’s a quote from a real climate scientist from the past,

    A fit to S(f) will be more weighted towards the low-frequency, high amplitude portion of the spectrum than a fit to log S(f). The latter procedure may provide a preferable alternative when the dynamic range of the spectrum is large. Any significance estimates that are strongly dependent on whether the red noise background is fit to S(f) or log S(f) should not be interpreted with great confidence.
    {appendix of Mann & Lees (1996)}

  14. farmerbraun says:

    Izen. Yep , all of that was explicit at the outset; it’s just a model right?
    The question is ; what can it do.

  15. izen says:

    “The question is ; what can it do.”

    Model = maps

    Perhaps it is possible to have a mathematical model based on how electrical energy behaves in complex circuits that would mimic climate behaviour sufficiently closely to derive some useful information about the way, and degree to which the climate can change. It matters that it makes testable predictions about measurable physical reality.

    Like the London tube map

    It would be useful for knowing how to get from one place to another, but provide no accurate insight into the physical reality above ground.
    It would be a useful description, but provide no explanatory value.

    The same applies to the Turing test chatbot. Although there the issue is a little deeper between how useful or real a mere mimicry of sentience might be.

  16. farmerbraun says:

    That’s pretty much it. It’s testable , we are told , and falsifiable as to accuracy. If it does work , (the claim is that it predicts cooling) we’ll still want to know why it works.
    It is interesting that there have been only a couple of minor quibbles ; “did you do this /did you do that ” type of stuff.

  17. karabar says:

    I find the discussion at JoNova quite interesting from a sociological standpoint.
    The process, I think, is operating quite well. David, and I presume Joanne, have been slaving away on this concept for a year and a half. It is at the stage now that it could be sent out for “peer review” and presumably published, assuming an editor could be persuaded to publish a paper that doesn’t scream “It’s that bloody CO2 wot did it”.
    This is the unfortunate situation that has evolved in the nonsensical field of so-called “climate science”. Apply for a grant on the basis that you will write a paper which illustrates that the evil capitalists are about to burn the house down, concoct an asinine paper based on imaginary data, circulate it amongst a few buddies for a rubber stamp, and find a publisher. Then hope that Steve MacIntyre and Ross McKitrick don’t discover your circular reasoning and other furphies and make you with draw it. (After having grabbed all the government grant loot of course).
    David’s work is at the stage for review, and essentially they are allowing the world to review it. I think there are sufficient knowledgeable folks that it is getting a thorough edit. There are, of course, a couple of masterful critics. Willis Eschenbach came out swinging, but seems have become more constructive lately. This is only natural. Willis has his own theory, and has a need to defend it against all comers. Lubos Motl has some constructive criticism, which, at first glance seems to be a bit antagonistic, but taking into account the Lubos is probably distracted by events in the Ukraine, a little English as a second language issue, and some academic opposition to a different process which is not part of his academic vocabulary, (not to mention that he is accustomed to being paid for his work) is contributing to some second thoughts on David’s part.
    All in all, I think the process of theoretical development, if that is what this is, is a fast track approach that might well become the norm. I find it similar to the problem-solving and decision-making processes adopted by successful organisations that seek employee involvement.
    The mainstream (i.e., Mann, Jones, Hansen, Patchy, Trenberth, etc.) has created a fantasy that there exists an imaginary phenomenon of a ‘warming’ which is in fact a ‘cooling’ which makes hindcasting with such a model virtually meaningless. What pre-1979 data could one use that has not been the victim of tampering? So, if the model comes together it will be at least a couple of years until we know whether or not the behaviour of the Evans ‘black box’ is representative. If it is, the task of determining why will be a challenge to say the least.

  18. karabar says:

    This comment from a TonyM sort of characterises what I was trying to say.

    “I guess this vindicates Jo and David’s idea to release it in sections to get feedback so that modifications can be made. I would like to add my congratulations and appreciation for their hard work and inviting us on their journey in exploring this hypothesis wherever it leads. David made it very clear from the outset that it is a hypothesis subject to testing along the lines of the scientific method; he is not wedded to it.”

  19. Ozboy says:

    I’m on the road at the moment, folks, but a brief observation on the events at Jo Nova’s:

    Jeez, it’s like a bloody cage match over there at the moment! Full-on donnybrook. David, Willis Eschenbach, Luboš Motl, Christopher Monckton, all swinging roundhouses! Back when I was a doorman, this is the point at which we washed our hands of the whole thing and called in the cops. I’m sure the warmists are making merry…

    I’m ashamed to admit it, but just at the moment the gory detail is right above my head. I’ve downloaded David’s paper on Optimal Fourier Transforms (OFTs), but it will be two weeks at least before I have time to get stuck into it. Understanding this technique is crucial to comprehending David’s new solar model. Fourier analysis is something I recall doing once at university, but haven’t looked at in the decades since. I had a couple of mates doing electrical engineering and they used Fourier transforms all the time.

    One high-level criticism I’ve noticed a bit of, is that the model does not have a great deal of explanatory value; that the “notch filter” is, as Izen observed, a post hoc bolt-on to make the model work (granted though, he admits up-front to adopting a black-box approach to irradiant mechanism, so underlying explanation isn’t the key aim). I’ll reserve judgement until I’ve had time to understand it more fully.

  20. farmerbraun says:

    Yeah nah, I reckon that it has been fairly restrained. Willis comes across as a hard man but he’s O.K. He just wanted to get stuck into it but he didn’t have the data . . yet.

    A fairly polished performance from the home team , I would say.

    I noticed Jo censored you earlier… how dare you criticise William Connolley on her blog! Oz 😉

  21. izen says:

    So we have a body of scientific knowledge developed over more than a century with significant progress in the last few decades in detecting and understanding the actual physical and chemical processes involved.
    Let’s call it the theory of evolution by natural selection.

    A small group with a history of antipathy to this theory independently develop a hypothesis about the path of evolution invoking irreducible complexity and a ‘black box’ out of which the observed variation emerges. Apart from minor traction with parts of the US educational system in the southern states this gets no traction in the scientific community which is ‘dogmatically fixated’ on the Darwinian explanation of the path of biology. Especially given the decades of discovery on DNA, genes and all the underlying complex hereditary machinery.

    I can’t help seeing a comparison between a century of development of understanding the climate, with recent advances in understanding radiative transfer, lapse rates and the complexity of the global circulation.

    And a small group with an antipathy towards the implications of this science developing a black box that provides an alternative explan- correction, description of the path of the climate.

    “Lubos Motl has some constructive criticism, which, at first glance seems to be a bit antagonistic, ”

    From what I have seen of Motl ‘antagonistic’ is just his usual approach for which he has become …er notorious? Take a look at his website, if that page design isn’t aggressive it is certainly challenging!

  22. karabar says:

    “century of development of understanding the climate”???????

    And they still haven’t figured it out!

  23. izen says:

    @- karabar
    “And they still haven’t figured it out!”

    There is still a lot of genetics that is uncertain at the bleeding edge, –

    Or, most of the complex interactions between genes act to constrain any possible variation. They increase the stability of the genotype by amplifying the damage local change may have.

    Or, group selection is not supported as a factor in the evolution of social traits. Kin selection and logistics are sufficient, and possibly necessary.

    But that does not refute or invalidate the stuff that is established like the basic Mendelian pattern of inheritance, or the role of genotype phenotype plasticity interactions on evolving populations.

    The two issues meet here….

    Or, the genotype change and phenotype plasticity in response to climate change can be measured in a number of plants. Some may not be able to keep up…

  24. karabar says:

    By George, Izen, you’ve got it! (finally)
    The climate is driven by teh sun.
    C02 has nothing to do with it.

  25. farmerbraun says:

    They’re getting a bit snotty with each other over here too. What’s up with that?

  26. izen says:

    “They’re getting a bit snotty with each other over here too. What’s up with that?”

    Probably the heat…

    @-karabar says:
    The climate is driven by teh sun.
    C02 has nothing to do with it.


  27. izen says:

    @- farmerbraun

    But, but they aren’t wearing any physics! ( clothes) grin!

  28. Ozboy says:

    G’day everyone,

    On the road at the moment and very little online time. If a comment doesn’t show up, don’t worry – I’ll attend to it within 24 hours or so.

  29. farmerbraun says:

    But Izen , the court of public opinion has no use for such high-flown notions. 🙂
    But seriously , if this model is shown to make valid predictions then the physicists can work out why. They may be some time.

    The important thing here is to provide politicians with an excuse to do what they already know that they must do; build more power generation , and the cleaner the better.

    And funnily enough that is precisely what most nations are doing.

  30. izen says:

    @- farmerbraun

    Here is the falsifiable prediction from the David Evans model –
    “Here’s the criterion: A fall of at least 0.1°C (on a 1-year smoothed basis) in global average surface air temperature over the next decade.”

    I am trying to decide whether this is incredibly ….’brave’ to predict a faster rate of falling temperatures than has been seen for a couple of centuries in the month which is not only the warmest in the historical record, but the first time there have been two consecutive months that break the record. And at the start of what is increasingly likely to be a strong El Niño.

    Or whether if this year and the first half of the next continue to break historical temperature records… And then 2017 is cooler with a return to La Niña, that is invoked as spurious ‘proof’ of the model.

  31. farmerbraun says:

    It’s a reasonable question Izen, but put it in the context of climate “science”.
    If 17 years of flat temperature is not falsification of models which predict steadily rising temperature, then what would constitute “spurious” proof? Or disproof?

    What we might have is a correlation , . . . or not.
    I’m not sure about your el Nino prediction. I have read a couple of items recently that suggest it will come to very little , and that la Nina will continue to dominate , as it has for the last 15 years or so.

  32. farmerbraun says:

    Izen , what do you reckon? I think that we may have a trend.
    It seems to Farmer Braun that several of the notables at WUWT, who coincidentally seem to have been working on climate models of their own, are experiencing the extreme discomfort caused by twisted nickers.

  33. karabar says:

    The symptoms most definitely point to a severe case of twisted nickers.
    On W. certainly displays all the typical signs of a jealousy attack, as I noted on JoNova a couple of weeks ago. He spit the dummy, collected his bat and ball and went home. But now on WUWT his home turf he has suddenly turned GREEN………..with envy I think.

  34. karabar says:

    Farmerbraun…….have you any idea who Rereke Whakaro is??? I know, just ’cause you are in Godzone doesn’t mean you know all 4 million. In any event, it woul dbe worth a trip over just to have a beer with the guy (and you). He is a character!

    G’day Karabar and everyone here,

    I’m finding it very frustrating that I have no time just at the moment to get my head around David Evans’ new model. I had hoped to digest it and provide some analysis on a new thread by now. I’ll be back in my office in about 3-4 days, and will try to find some time after that; I hope the Excel sheets will be available by then too.

    One quick observation though, and it picks up on Izen’s point – the 0.1K criterion. It does appear the sort of thing that could easily be drowned out by the noise of more proximate factors, including El Niños and La Niñas, the PDO, et cetera. But there you go: he’s put it out there, and laid his theory open to falsification, as any good scientist should. The time scale is a bit of a concern; the fall is supposed to begin any time between now and 2024, though his reasoning behind this range appears sound to me. So I guess it’s now a waiting game – Oz

  35. karabar says:

    This is where David explains that for you
    David Evans
    June 28, 2014 at 2:55 pm · Reply
    Dikran: The graph shows 0.5C or more, but as explained in the article the land thermometer data the solar model trained its parameters on might be exaggerated — in which case the solar model in the diagram would be too sensitive and it would show a fall that is too large. (If it learned to match rises that were exaggerated, its predictions of future falls and rises would also be exaggerated.) So may 0.25C.

    Make an allowance for some CO2 warming, maybe 0.05 over a decade say.

    Hence a cautious estimate of a 0.2C fall.

    CO2 models say 0.2C up, but maybe a continuing pause — anything but down. So a criterion of 0,1C would differentiate the models nicely.

    The model results shown are 1-year smoothed, which is enough to knock out the monthly fluctuations and weird spikes that may well be measurement errors. So I went with that.

    The decade starts when the article was published, 27 June 2014.

    Not interested in the average of the decade, because it might get warmer before it gets cooler (there are reports of a developing El Nino). Just where it ends up after a decade.

    Although I assume it’s either CO2 or notch-delay solar in some parts of the series, the criterion here is good enough to falsify either of these theories even if there is a third explanation.

  36. Ozboy says:

    A brief OT rant…

    Jeez the Rolf verdict leaves me gutted! Back in the 60s, when as a young kid I watched him on TV, I thought there was something seriously creepy about him. And told my mum so. I always suspected him of being some kind of paedo, simply on the basis of my gut reaction to him. Having that suspicion vindicated offers me no joy at all. Bugger it, he was a world-famous Aussie: an ambassador for me!!! And he was a @#$%ing rock spider all along. Larry Pickering knew about him all along.

    I feel sorry for his long-suffering family. May he rot in jail.

  37. farmerbraun says:

    One comment from above link:-

    “I feel quite sorry for sad old groper Rolf Harris, but this is payback for those terrible television programmes I was forced to watch while my boyfriend and I waited for his parents to go to bed so we could fool around.”

    A rather disturbing theme that emerges from a lot of the articles on this is “oh, he wasn’t a paedophile, he was just a dirty old man who groped young girls”. Say what? A fifty-something entertainer screwing his thirteen-year-old daughter’s bestie while his wife slept in the next room??? What I don’t get is why one or more of the dads didn’t escort him out the back for a session of, um, gentle instruction and encouragement.

    You might think that’s over the top. But the essence of Libertarianism is a mutual regard for the liberty of others. He was bloody adept at taking liberties, all right. But he gave not a rat’s arse about anyone else but himself – Oz

  38. farmerbraun says:

    I don’t disagree Oz. But bigger rats may be flourishing.

    Oh, yes. Much, much bigger – Oz

  39. izen says:

    @- Ozboy
    Like this ?

    Lord Brittan has confirmed he received a ‘substantial bundle of papers’ from MP Geoffrey Dickens in 1983 – which detailed allegations of a paedophile network within Parliament and Whitehall – and passed them to his officials for investigation.
    But amid claims of an Establishment cover-up, the Home Office admits the dossier was either lost or destroyed.
    …A child abuse campaigner who worked closely with Mr Dickens said the missing dossier could have toppled the government of the day….
    But she claimed some belonged to Mrs Thatcher’s government – and called on David Cameron to help end the cover-up. ‘This file will never see the light of day because it could have brought a government down. It will have been shredded and burned,’ she said.

    My reaction to it is the same irrespective of the political parties involved. I’m a dad. For me, it’s personal, and Thatcher’s government be damned. Fiat justitia ruat caelum – Oz

  40. Ozboy says:

    David Evans has his Excel spreadsheet up with all data and code visible. I’ve downloaded it. But I’m drowning at the moment – it’ll be weeks before I can do anything with it. I might try to get a new thread up next week.

  41. Ozboy says:

    Well I normally don’t watch soccer, but I got up at 0500 here to watch Brazil play Germany in the World Cup semi-final. At least I can tell my grandkids I saw it live. Bloody hell!!!

    The word blitzkrieg does spring to mind… Deutschland über alles!

Comments are closed.