Slashdot Mirror

← Back to Stories (view on slashdot.org)

Breakthrough In Automatic Handwritten Character Recognition Sans Deep Learning (technologyreview.com)

Posted by Soulskill on from the bet-it-wouldn't-work-on-mine dept.

subh_arya writes: Researchers from NYU, UToronto and MIT have come up with a technique that captures human learning abilities for a large class of simple visual concepts to recognize handwritten characters from World's Alphabet. Their computational model (abstract) represents concepts as simple programs that best explain observed examples under a Bayesian criterion. Unlike recent deep learning approaches that require thousands of examples to train an efficient model, their model can achieve human-level performance with only one example. Additionally, the authors present several "visual Turing tests" probing the model's creative generalization abilities, which in many cases are indistinguishable from human behavior.

66 comments

  1. #gamergate by Anonymous Coward · · Score: 0

    i like video games

  2. human-level performance by Anonymous Coward · · Score: 2, Funny

    their model can achieve human-level performance with only one example

    Yeah? Well, I've never encountered a human - myself included! - who can read my handwriting, so suck it, you AI mofos!

    1. Re:human-level performance by davester666 · · Score: 2

      Now all doctors will have to go back to school and learn how to write even worse than they do now. I presume it'll be something along the lines of:

      -just scribble for 3-4" on the paper, using a felt marker, making no attempt to actually move in the shape of any known letter
      -wet tip of finger
      -rub the scribble for several seconds

      --
      Sleep your way to a whiter smile...date a dentist!
    2. Re:human-level performance by Chrisq · · Score: 1

      Yeah? Well, I've never encountered a human - myself included! - who can read my handwriting, so suck it, you AI mofos!

      Some people are incredibly good at this. My wife used to type up dissertations in the days before students did their own on computers, and she can read things that to me are completely illegible. I thought my handwriting was bad but she says she has seen much worse.

      Of course the question is, are they aiming for average human ability or someone who is practiced in reading difficult handwriting?

    3. Re:human-level performance by Anonymous Coward · · Score: 0

      Hell, I've run into plenty of captchas that I couldn't read, and they're not even hand written, so I hope the AI is better than me.

  3. Breakthrough saving humanity! by Anonymous Coward · · Score: -1

    ISIS is bombing passenger planes, gunning down crowds of civilians, burning people alive in metal cages but this is all you nerds can think about. God I hope there's a war draft to get all you basement dwellers out front and center with the cold reality this nation faces. You cant build yourself a firewall from ISIS. You need guns, bombs and a pair o boots for some sweat equity in your nation's future survival. Sorry, nerds but sweat from masturbation does not count!!

    1. Re:Breakthrough saving humanity! by quax · · Score: 0, Offtopic

      This is much more important than Daesh/ISIS. If you think the US is threatened by these barbarians then they already achieved half their objective, getting into your head and under your skin. The other half of their objective is to get US troops involved into another mid-East quagmire.

    2. Re:Breakthrough saving humanity! by Zontar+The+Mindless · · Score: 1

      This is probably the same AC who complains every time a story mentions any developments relating to ISIS that it's not "news for nerds".

      --
      Il n'y a pas de Planet B.
    3. Re:Breakthrough saving humanity! by gtall · · Score: 1, Offtopic

      I know I shouldn't respond to a troll, but you and congress-critters need to stop having wet dreams over a "draft". Modern warfare by the West is a professional occupation and requires much training. The unprofessional rabble that Iraq and Afghanistan's governments field show just how ineffective untrained meat-sacks are against religious zealots. Even Iranian backed Iraqi militias are pretty much cannon fodder with loud mouths. Ever watch the Kurds fight? They are just a bunch of guys with guns.

      The U.S. cut through just about anything the Al Qaeda could field. Where it had problems was in attempting to solve the underlying issues that spawned the problems starting with Muslim 7th century wet dreams. One can see just how removed the current populace's idea of modern warfare is by listening to their mouthpieces in Congress. McCain (who ought to know better but stupidly doesn't) made some remark about how carpet bombing Germany finished off the Nazies. It didn't, Hitler was a more effective weapon against the Nazies by being a dolt. What carpet bombing did do was cause a massive human tragedy which put a big dent in the Marshall plan after the war. Cruz is another moron advocating this.

      When asked if the U.S. would carpet bomb ISIS, Gen. Paul Selva (vice chairman of the Joint Chiefs of Staff) in front of McCain said unequivocally, "The U.S. Military will never carpet bomb." He and the U.S. Military would be brought up on war crimes in EuroLand. But that isn't why he dismissed such a stupid move.

      While we can be dismissive of that lot over there, public opinion matters, the general and the U.S. Military knows that the U.S. and the West is in an ideological war with Islam, to treat it as purely kinetic is just mindless. Europe, whatever their faults, did the Christian thing and accepted untold thousands of refugees from Syria but also Afghanistan, and N. Africa. They put the "Christians" in the U.S. to shame. Even Little Jerry Falwell at Liberty U. was counseling his students to arm themselves against the Muslims. I guess Jesus had a secret directive, "Turn the other cheek that I may get a better aim at your ass."

      The general and the U.S. Military know damn well that the U.S. needs to work on the underlying problems that carpet bombing will not solve. It would only antagonize the pop., something Russia is conveniently forgetting, and not stopping Russia from doing this by giving Putin a swift kick in the nuts will cause spill over that the West will suffer from for years.

    4. Re:Breakthrough saving humanity! by brianerst · · Score: 1

      I read your comment about congress-critters having "wet dreams over a draft" and assumed you'd be talking about Democrats, but, oddly, you seem to think Republicans want a draft. The only bills introduced in the last decade-plus to reinstate the draft were sponsored by Democrats (Charlie Rangel seem to bring one up every once in a while). Left leaning policy wonks have brought up the idea, and Dana Milbank famously wrote an op-ed promoting a new draft.

      Now, some of this is basically an attempt to highlight alleged hypocrisy or to "spread the pain" outside of the perceived "only poor folk join the military" viewpoint of many on the left. Leaving aside that volunteering for the military isn't particularly clustered around poverty (it is clustered around a history of family members in the military), this is pretty cynical fear-mongering that's not much better than what the Republicans do. And to the extent that it's sincerely held belief (e.g., Jim McDermott, (D) WA, who honestly believes in a draft), it's a pretty daft one that generally revolves around social engineering (forming a "more engaged" populace through compulsory service).

      Honestly, neither party is all that coherent when it comes to foreign relations and the military. Both sides are happy to bomb brown people all over the world and both want to do it on the cheap. It's a choice between feckless and reckless.

    5. Re:Breakthrough saving humanity! by quax · · Score: 1

      You are probably right :-)

  4. Timely discover considering nobody writes anymore by JoeyRox · · Score: 5, Funny

    Maybe they'll also invent a better way to untangle corded phone cables.

  5. Re:Timely discover considering nobody writes anymo by Anonymous Coward · · Score: 1

    You do that by euthanizing old people.

  6. thousands of examples to train an efficient model? by Anonymous Coward · · Score: 0

    You sound like my kids.

    1000s of examples to train? Boo hoo. What's that... A few hounded nanosecond?

    Fuck me.

  7. Great by sexconker · · Score: 3, Insightful

    I'll never solve the new captchas.

    1. Re:Great by locopuyo · · Score: 2

      I just use a browser plugin for it.

      --
      The Official Site of 1337 Pwnage
  8. Well that's cleverer by Zero__Kelvin · · Score: 3, Insightful
    Seriously? Cleverer? Oodles ? What editor left those in the paper? Slashdot editors must be working for these guys on the side. I know somebody will say cleverer is technically correct, and while that may be true, it is a disaster aesthetically.

    "“The real inflection point in AI is going to come when machines can actually understand language,” he says. “Not just doing mediocre translations, but really understanding what you mean.”"

    Until we understand what it means to understand, how can we possibly know if we have taught these systems to understand? Even if it responds intelligibly, and what it says makes complete sense, is that the same as understanding? I suppose as Billy C. once said: "It depends on what the meaning of the word 'is' is".

    --
    Guns don't kill people; Physics kills people! - John Lithgow as Dick Solomon on Third Rock From The Sun
    1. Re:Well that's cleverer by Anonymous Coward · · Score: 0

      Well look at mister clevererly.

    2. Re:Well that's cleverer by phantomfive · · Score: 1

      Until we understand what it means to understand, how can we possibly know if we have taught these systems to understand?

      If I am talking to you, I can generally tell if you understood what I said or not, even if the meaning can't be clearly defined. Presumably an AI will respond similarly, and I'll be able to tell if it understood or not.

      If we get AI that smart, then we will have advanced a long way.

      --
      "First they came for the slanderers and i said nothing."
    3. Re: Well that's cleverer by Anonymous Coward · · Score: 0

      I're the mostest clevererest!!

    4. Re:Well that's cleverer by Prune · · Score: 1

      Even if it responds intelligibly, and what it says makes complete sense, is that the same as understanding?

      Good job poorly rehashing an argument made in 1980

      which has been refuted time and again.

      --
      "Politicians and diapers must be changed often, and for the same reason."
    5. Re:Well that's cleverer by Anonymous Coward · · Score: 0

      Faggot.

    6. Re:Well that's cleverer by HiThere · · Score: 1

      Unfortunately, the arguments are not conclusive. It really does depend on the meaning of understanding. It's also, however, true that the presumption of lack of understanding isn't defensible.

      Lacking an exact definition of understanding, the only thing we have to go on is something like "Well, if I had reacted that way, then it would mean that I understood.". This is clearly inadequate as a non-observer relative description...something which even quantum physics manages to come up with, though it puts limits on the precision of possible observations.

      This is a real problem as when there is no observer-independent evaluation of the ability of a system to understand, then it's quite easy to claim that it doesn't when there is any advantage to doing so. (People have even used this argument against other people.)

      Another problem is that without a good measure of understanding, we confuse different models of the universe with lack of understanding. This is readily observable in any heated discussion, and generally leads to ill-will on both sides.

      It is quite plausible that even given an exact definition of understanding we would not be able to apply it to mammals, insects, etc. as it may require a precise description of parts of the thought process, but it should, in principle, be possible to apply any exact definition to any artificial intelligence. Do note, however, that I said "in principle". This is because it's already true that for many AI programs developed using variations of evolutionary techniques nobody actually understands how and why they are working as they do.

      --

      I think we've pushed this "anyone can grow up to be president" thing too far.
    7. Re:Well that's cleverer by azcoyote · · Score: 1

      The viability of the "-er" comparative ending in English varies from place to place; in Canada, for example, more "-er" words are acceptable than here in the United States (e.g. "funnier" I believe works in Canada). This is never merely a matter of what is technically correct, however, because our aesthetic aversion to this or that form is already determined beforehand by common practice, such that I feel that "cleverer" is awkward simply because it is not proper in the USA. If I had grown up in Canada, my very aesthetic sense would likely be different.

      --
      Incipiamus, fratres, servire Domino Deo, quia hucusque vix vel parum in nullo profecimus.
    8. Re:Well that's cleverer by Zero__Kelvin · · Score: 1

      "If I am talking to you, I can generally tell if you understood what I said or not, even if the meaning can't be clearly defined."

      Really? What does it mean to understand, since you seem to be the only person I know of who claims they can answer that question? Your reply shows that you didn't understand what I wrote at all; ironic, isn't it ;-)

      --
      Guns don't kill people; Physics kills people! - John Lithgow as Dick Solomon on Third Rock From The Sun
    9. Re:Well that's cleverer by Zero__Kelvin · · Score: 1
      Are you some kind of idiot? Because someone else also understood a problem in 1980 I have no right to talk about it now? I wasn't aware there was a huge checklist of topics, and that each time someone covers one nobody else can ever understand it or talk about it again.

      which has been refuted time and again.

      BTW, it was a rhetorical question; you clearly are an idiot. There is nothing to refute in anything I wrote in my original post. It was a series of questions. How exactly do you refute a series on non-rhetorical questions anyway there McFly?

      --
      Guns don't kill people; Physics kills people! - John Lithgow as Dick Solomon on Third Rock From The Sun
    10. Re:Well that's cleverer by phantomfive · · Score: 1

      There may be a grey area, certainly, where it's hard to tell if something understands or not, but for today's AI, we can certainly say they don't understand (unless you come up with some narrow definition of "understand")

      --
      "First they came for the slanderers and i said nothing."
    11. Re:Well that's cleverer by Zero__Kelvin · · Score: 1

      Holy shit. This is frigging hilarious. Seriously. You literally don't understand that defining what it means to understand is something nobody can possibly understand. Priceless.

      --
      Guns don't kill people; Physics kills people! - John Lithgow as Dick Solomon on Third Rock From The Sun
    12. Re:Well that's cleverer by phantomfive · · Score: 1

      That makes sense.

      --
      "First they came for the slanderers and i said nothing."
    13. Re:Well that's cleverer by Zero__Kelvin · · Score: 1

      It's pretty clear that you don't understand.

      --
      Guns don't kill people; Physics kills people! - John Lithgow as Dick Solomon on Third Rock From The Sun
    14. Re:Well that's cleverer by Zero__Kelvin · · Score: 1

      Nope. Cleverer is clumsy, and your insinuation that Canadians are naturally clumsy is offensive ;-)

      --
      Guns don't kill people; Physics kills people! - John Lithgow as Dick Solomon on Third Rock From The Sun
    15. Re:Well that's cleverer by Prune · · Score: 1

      How exactly do you refute a series on non-rhetorical questions anyway

      Because, while I might (for the sake of argument) accept they're not outright rhetorical questions, they're also certainly not questions made in good faith -- they're implying that one can reasonably suppose there's a possibility of a difference between understanding and the functional competence exhibited by an entity that understands (equivalently, that, above some threshold, the appearance of intelligence can possibly be different from actual intelligence -- and feel free to replace "intelligence" with "understanding" and "consciousness" -- for while they're not the same things, they're related and the same class of arguments apply that such a difference is fantasy).

      --
      "Politicians and diapers must be changed often, and for the same reason."
    16. Re:Well that's cleverer by Prune · · Score: 1

      It really does depend on the meaning of understanding.

      This is the same sort of sophistry that philosophers used to engage in when discussing qualia, until Dennett showed it was all bullshit and they're just emergent ephiphenomena.

      Understanding just means a sufficient level of integration of some information with knowledge already extant in your mind -- the various semantic elements of what you understand are linked to the rest of your knowledge, so that you can relate these and also make use of the new information as a model of the target of your understanding (where most of the rules for emulating the model in your head derive from the analogies and relations to your previous knowledge). Said integration can occur to varying degrees, but that's equivalent to saying understanding can be shallower or deeper, which no one will disagree with.

      --
      "Politicians and diapers must be changed often, and for the same reason."
    17. Re:Well that's cleverer by Zero__Kelvin · · Score: 1

      The problem you have is that you have made my point while attempting to refute it :-)

      --
      Guns don't kill people; Physics kills people! - John Lithgow as Dick Solomon on Third Rock From The Sun
    18. Re:Well that's cleverer by phantomfive · · Score: 1

      You are able to determine that.

      --
      "First they came for the slanderers and i said nothing."
    19. Re:Well that's cleverer by Zero__Kelvin · · Score: 1

      Yes, moron. It is much easier to show that something isn't understood, then that it is. Indeed, the only things we can determine are that they don't understand, or that they might understand, but never that they actually do understand. Off you go now little retard ...

      --
      Guns don't kill people; Physics kills people! - John Lithgow as Dick Solomon on Third Rock From The Sun
    20. Re:Well that's cleverer by phantomfive · · Score: 1

      You're such a pleasant chap. I'm still quite sure you understand.

      --
      "First they came for the slanderers and i said nothing."
    21. Re:Well that's cleverer by Anonymous Coward · · Score: 0

      It really does depend on the meaning of understanding.

      This is the same sort of sophistry that philosophers used to engage in when discussing qualia, until Dennett showed it was all bullshit and they're just emergent ephiphenomena.

      Understanding just means a sufficient level of integration of some information with knowledge already extant in your mind -- the various semantic elements of what you understand are linked to the rest of your knowledge, so that you can relate these and also make use of the new information as a model of the target of your understanding (where most of the rules for emulating the model in your head derive from the analogies and relations to your previous knowledge). Said integration can occur to varying degrees, but that's equivalent to saying understanding can be shallower or deeper, which no one will disagree with.

      Interesting point!

      From my point of view as an AI researcher and computer scientist as a career, There is something to be said about knowing when and when not to disregard demarcations between different domains of knowledge. This has a profound effect on the question of this thing called "understanding".

      From a computer science and engineering perspective, it depends on the use case of the problem whether 'understanding' of the device or program as an actor is sufficient to operate at a high enough level of accuracy and reliability to meet the requirements of the application being pursued. From a philosophical standpoint the question of whether a program "understands" the application is largely meaningless.

      Take the example of Deep Blue vs. Gary Kasparov. Deep blue was nothing more than a massively parallel database application that compared chess board configurations of pieces and graded them by different strengths for each player. In the use case of a computer versus a human chess game, Deep blue was massively successful at being a functional equivalent of a grand master chess player. Deep Blue had no concept that it was playing a game, it did not know who it was playing against, where or even when.. so the question of it "understanding" that it was playing a game would yield an answer of definitely not.

      There are a lot of insightful posts here dancing around this point on the question of understanding and what is meant by it as a general constraint of measuring the strength of an AI application: There is no over-arching definition of what, in a nuts and bolts way, what our wetware is doing to turn sensory input into actionable intelligence about the world that can be recognized, remembered and automatically responded to in such a way that we call "intelligent behavior". It is even to some researchers, seen as not a useful avenue of questioning, I think because of the domains of applications and questions being poorly defined.

      This does not mean that the question itself is meaningless, but that it must be approached carefully in terms of constraints and context.

      In the case of machine perception, if asked to define understanding, I agree with Jeff Hawkins, the author of "On Intelligence", that Understanding can be thought of as the ability of an entity to observe a phenomena in such a way that, in terms of pattern recognition and time it can observe a pattern or sequence of patterns and at an instantaneous point in time be able to predict what partial pattern, pattern or sequence of patterns is coming next. If I am sitting in a restaurant, and I say to you "Will you pass me the ..." you have a strong subconscious pull, rather automatically to respond with something like "Salt" or "Ketchup", depending on the restaurant. You might be surprised if I end that sentence with something unexpected like "Sidewalk" or "Sarcasm".. and you would as a human, automatically mentally back up and re-evalute what it is that I am asking or trying to express. Interestingly as a tangental example of what I mean by Use-case being relevant or from the perspective of anoth

    22. Re:Well that's cleverer by Prune · · Score: 1

      Do you often find that some things are obvious to you but others just don't understand? Of course, the most effective trolls are the ones that actually believe what they're purveying. Uncle Al in the newsgroups comes to mind...

      --
      "Politicians and diapers must be changed often, and for the same reason."
  9. Improvements to OCR? by pipedwho · · Score: 4, Interesting

    I hope this heralds in some significant improvements to basic OCR. It amazes me that OCR against a printed document still doesn't always yield 100% success. Even worse are OCRs on printed music manuscripts. The recognition and transcription quality is atrocious.

    And yet, these guys can recognise handwriting with incredible accuracy.

    I keenly await when these algorithms can be expanded to general OCR / document recognition. Even if there need to be specific models for each type of document.

    1. Re:Improvements to OCR? by Anonymous Coward · · Score: 0
      I don't know anything about this breakthrough, but my old galaxy note II could understand my Cyrillic handwriting pretty well (It was faster to use it than to poke the letters (swipe was not available)). And please note, that I am a terrible hand-writer, most of the time even I cannot read my writings.

      If you have the stroke information it is relatively easy to recognize. If you try to OCR it would be much more difficult to recognize handwriting. I also thought this was about OCR, because thought gesture based recognition was already done.

    2. Re:Improvements to OCR? by Anonymous Coward · · Score: 0

      It amazes me that OCR against a printed document still doesn't always yield 100% success.

      It might be partly because the OCR systems have to cope with too much information, and they do not resolve things like humans would. If you know 100 glyphs that look like the glyph on the scanned paper, it is difficult to pick the right one. Also, even if the scanned image is good, the OCR system still needs to assume there are errors in the scan as it cannot know it for sure.

      Usually the errors made by AI systems are pretty logical if you have the same information (both training data and hard-coded rules) as the AI system has.

    3. Re:Improvements to OCR? by Richard+Kirk · · Score: 3, Interesting

      Suppose you had a bit of your handwriting that you could not read. How do you figure out what you wrote. One thing that I do, and you may do too, is to try and imagine writing the thing, and work out the rhythm of what you are writing. If you can get some sense of how your hand is writing, you may see that what was a 'u', or maybe an 'n' or half of am 'm' makes sense because of the way it joins up to other stuff. We seem to have some sort of kinematic two-and-a-half axis model for writing. We use different muscles if we are writing with a pen (fingers and wrist), a blackboard (wrist and upper arm), a spray-can (upper and lower arm), or a tiny engraving tool (just fingers) and yet our handwriting remains much the same. So some computer that can try and fit the same kinematic model should make better guesses for a word it has not met before than anything that just trained on the shape.

      This does not directly transfer to OCR. If you have a page of fixed-width text, then every letter has its own little rectangle, and you can either recognize that using the traditional OCR model, or you can't. However, there is something we can do along the same lines. Suppose you have a document that you guess was rendered from PostScript. If you have a guess for a particular word, and the font it was rendered in; you could render that part of text. You can then degrade that rendered image to mimic the properties of the printing and scanning, and check the fit. The best solution will probably be the one that achieves the best fit with the shortest, and hence most probable bit of PostScript. When you have more text, you can pick up hints from the spacing, the justification, and other larger page layout structures.

      I actually worked on OCR, and tried both of these once. It might have worked with a large software team, but I hadn't got one.

    4. Re:Improvements to OCR? by Big+Hairy+Ian · · Score: 1

      Much as I agree this could herald a much needed improvement in OCR the downside is that Capcha will become virtually useless.

      --

      Build a Man a Fire, and He'll Be Warm for a Day. Set a Man on Fire, and He'll Be Warm for the Rest of His Life.

  10. Not just for culottes anymore by Anonymous Coward · · Score: 1

    n/t

  11. Yet another sign that std neural nets are dead-end by Anonymous Coward · · Score: 0

    This is yet another sign that current artificial neural networks are not the path to either understanding or implementing human intelligence, either qualitatively or performance-wise. Previous relevant Slashdot threads: [1] [2] I really don't understand why neural networks are still taught as part of AI courses, rather than say algorithmic courses, since neural networks, at least the way they currently are, have about as much to do with machine intelligence as quicksort.

  12. Considering nobody writes cursive anymore by littlewink · · Score: 1

    FTFY. BTW only cretans cannot write cursive.

    1. Re:Considering nobody writes cursive anymore by richy+freeway · · Score: 5, Funny

      It's cretin, you cretin.

    2. Re:Considering nobody writes cursive anymore by Anonymous Coward · · Score: 1

      Crete, the island in Greece, is so backward!

    3. Re:Considering nobody writes cursive anymore by Anonymous Coward · · Score: 0

      Are you serious? Most Linear B texts look pretty cursive to me.

  13. Oblig link: source code by Anonymous Coward · · Score: 2, Informative

    If only the /. editors would do some minimal investigation... Oh wait, this is still /.

    https://github.com/brendenlake/BPL

    1. Re:Oblig link: source code by grilled-cheese · · Score: 1

      If only the /. editors would do some minimal investigation... Oh wait, this is still /.

      https://github.com/brendenlake/BPL

      At the moment the code requires both Matlab and Lightspeed. Until someone ports the code to an OSS library or alternate language, it won't see significant adoption.

  14. Re: thousands of examples to train an efficient mo by Anonymous Coward · · Score: 0

    collecting samples is not always easy.

  15. Re: thousands of examples to train an efficient mo by Anonymous Coward · · Score: 1

    Or more: labeling them...

    RBMs/autoencoders as unsupervised learning algorithms are normally used for pre-training in semi-supervised learning.

    Maybe someone at google has made some breakthroughs in unsupervised learning but AFAIK, most of the deep learning scene still requires hundreds of thousands of labeled sample images.

  16. the ai last heard muttering by Anonymous Coward · · Score: 0

    "... that's definitely a d... or maybe a 6... th?..."

  17. EnduranceRobots.com is looking for enthusiasts by endurancerobot · · Score: -1, Offtopic

    EnduranceRobots.com is looking for enthusiasts and hobbiests in robotics and laser industry. We are looking for tech smart people who would like to help us to improve our products and positioning on the markets. We are still very early startup and can not pay big salaries but we can pay some. Please have a look our web site: endurancerobots.com youtube channels: http://www.youtube.com/channel... http://www.youtube.com/channel... and our facebook: https://www.facebook.com/Endur... We are very open to all critics: email us your ideas to: gf@endurancerobots.com or skype my: george.fomitchev

    1. Re:EnduranceRobots.com is looking for enthusiasts by Anonymous Coward · · Score: 0

      Skype: No way! That's nonfree (freedom-denying) software. Use Ekiga or an ordinary phone call

  18. Re:Yet another sign that std neural nets are dead- by byrtolet · · Score: 1

    I really don't understand why neural networks are still taught as part of AI courses...

    Because they have to teach something.

    Since no AI exists (yet), I've always thought this is a stupid subject name!

  19. Oh the spam. by Anonymous Coward · · Score: 0

    Someone broke captchas?

  20. Comparing apples to oranges by Anonymous Coward · · Score: 1

    This is just clever way to describe features on the images. They have built a codec for hand-writing. Instead of a binary array you are learning a set of brush-strokes. I bet that if deep learning algorithm is given the decoded brushstrokes as learning material, it will outperform this.

    1. Re:Comparing apples to oranges by mrego · · Score: 1

      What of calligraphy?

  21. Links: paper and video interview w/ author by RandCraw · · Score: 1

    The paper:
    https://www.sciencemag.org/con...

    A short article and interview with Lake:
    http://www.ibtimes.com/say-hel...

  22. The usual pattern for AI research by Anonymous Coward · · Score: 0

    An advance is announced. It is advertised as a breakthrough. It works impressively in some specific examples. It is then applied to more general examples. Unexpected difficulties pop up. The performance is not bad, but only incrementally better than that of previous solutions. Researchers move on to a different paradigm, until the next hype.

  23. parsing ridicule by epine · · Score: 1

    It depends on what the meaning of the word 'is' is

    This has become one of the best memes ever to sort the kind of people who just snort and move on in the rush to stop thinking from those of use who perceive Clinton's comical perch as resting upon a legitimate labyrinth of linguistic complexity.

    The same impatient mind is at some point informed that the Chinese language has no tense system as we know it from most European languages. "How does that even work?" these people ponder for a few tense milliseconds, before it gets filed under "Stonehenge" and/or epicanthic enigmas.

    It's not nearly so enigmatic as all that. Even in English with formal tense markers, we navigate our tense boundaries with incredible subtlety. When precisely does the present become the past? Entire SF novels have been authored to address this theme, not to mention a parallel, button-down literature bereft of "oodles".

    It's only in a game of persecutorial small minds (Kenneth Starr you are) when one compels one's counterparts to answer questions in the bullshit Booleoverse that our customary subtleties run around with their flies open.

    "Yes or no? Is there a sexual relationship between you and the celebrity intern in the blue dress?"

    If you're not living in the bullshit Booleoverse, the conversation goes like this:

    Q: Is there a sexual relationship between you and the celebrity intern in the blue dress?

    A: Not presently.

    There, that's the kind of outrageous subtlety those clever Chinamen (and Chinawomen) employ all farming day with their "tenseless" language.

    "Oodles" is actually the worse problem, because the chinaman/women (if only there was a language which didn't compel gender) who studies ESL diligently all through their schooling might very well not know the word. These same people adore "cleverer" and wish the entire damn English language would be half so consistent. One man's euphony is another man's cognitive burden.

    What Clinton should have answered: "It depends on what kind of fatuous asshole is asking the question." But that might have come across in a negative way. In trying to choose the lesser evil, Clinton ends up painting himself into a comical corner, reminding me of how Ainsley Haynes once tried to relieve her burbling bladder in the Oval Office coat closet.

    In a rational world, the whole thing should have been framed as a debate about the precise location of the tumour on Clinton's moral spinal cord (the tumour itself was never in doubt). Was his disorderly conduct confined to regions below the belt only, or did it extend to his Presidential Seal and signing authority on the wrong and unacceptable end of the Nixon--JFK misbehaviour spectrum?

    The Republicans, determined to howl equally loudly either way, were trying to assert that his "lack of judgement" in wiener deportment made him unfit for leadership (which would come [name that tense] as a shock to half the great men of history).

    The spectacle's circular logic boiled down to this:

    "If the man can't figure out that the Puritan sentiment in American will allow us to hound him over this incident to the four corners of the earth assuming the anti-intellectual posture of fatuous assholes (which hardly anyone outside of France will call into question), he's not presidential material in the first place; cogito argal sum we should bust his balls for the good of America by any means necessary."

    Many frame the issue as one of lying in office, which I suppose it was. However, by that standard one would have to rate ballot manipulation in Florida to gain office as a RICO version of the same offense.

    To further buffer the bullshit, Christians apparently have a special door for the purpose of turning "is" into "was": accepting Jesus Christ as the one true saviour. "I was a sinner, but now am saved."

    Now in the mind of Kenneth Starr and his like-minded brethren, BC sure as hell wasn't loudly beating

    1. Re: parsing ridicule by Anonymous Coward · · Score: 0

      Your prose style shows promise but is too rococo for Slashdot.

      Clinton should of course just have said, "Yeah, Monica sucked my cock. So what?"

  24. Re:Yet another sign that std neural nets are dead- by HiThere · · Score: 1

    You are being overly simplistic. Some parts of an AI will use elegant heuristics to solve certain classes of problems. Others use other a tangled mass of spaghetti, where no efficient approach has been discovered. And there will be LOTS of different tools. Think of most of an AI as being analogous to a code library. That's still overly simplistic, but it addresses this point. There will be specialized pieces that are optimized to handle certain classes of problem. There will be other places where you need to cobble together something that will work.

    One thing that makes the above paragraph overly simplistic is that there is no one central "main" routine. I've become rather convinced that consciousness is the result of the serialization necessary for coded memory retrieval, and that language is a very late and minor routine developed from that purpose. The underlying thought processes are incredibly parallel, so most of it is pruned away in the building of indexes...though when you retrieve the referenced memories (NOT original, and mutable whenever accessed) you are able to retrieve many of the parallel threads, and occasionally even ones that are not sensation based. (Language memories are generally either memories of motion of the vocal cords or of the auditory sensation, and thus qualify as sensation based.)

    --

    I think we've pushed this "anyone can grow up to be president" thing too far.