Skip to main content

I'm sure that many of you watched two of Jeopardy's greatest living human players go up against Watson last night. So far, there's been a lot of differing opinions on it. Some are awed by the ability of a computer to go up against humans, in their own language, and hold its own. Others shrug and say "Get back to me when it can do something useful".

For those of you that aren't in awe, you probably should be. And those of you that are in awe, may be in awe for the wrong reasons.

Follow me after the jump, and I'll explain why.

To begin with, Watson's a pretty impressive piece of machinery. Or, more specifically, a cluster of them. It's 100 Power 750 servers, full loaded and decked out. That, plus the racks, the networking gear to tie it all together, the additional storage, etc. -- you're looking at $20 million to $24 million worth of hardware. If not more. It has some amazing computational power -- it's an absolutely massive database that it has to wrangle (and it's why humans skunk it on questions that take Alex Trebek a short time to read -- it isn't quite real-time yet).

But Watson's real promise isn't the fact that it can play Jeopardy. That's simply the flash to raise interest in the project, and it also gave the brilliant minds behind the project some focus. As anyone who has decided they want to learn to program or play an instrument knows, it's a lot easier to ,ale "I want to make a program that does this" or "I want to learn how to play Hotel California on the guitar" happen than "I want to learn to program" or "I want to play guitar".

Its real promise isn't even artificial intelligence. And if you're in awe of it because it seems like it is intelligent, or close to it, you're in awe of it for the wrong reason. Watson isn't intelligent, and isn't anything close to intelligent. Your dog is more intelligent than Watson is.

Watson is a machine designed to be good at one thing -- analyzing variables in context given a dataset of nearly infinite variables and contexts. And if that sounds like a downer, it isn't. That is something to be in awe of.

Right now, Watson gets the clue fed to it in a text file. You could make it listen and analyze sound, but you'd be massively increasing the cost, and as a proof of concept, it's worthless. We already know we can analyze sound and teach a computer what to do with that sound. That's why you have voice-activated music players in case, voice dialing on phones, speech-to-text software, etc. It's flashy, but useless for what Watson is showing off.

It also doesn't see anything, or more specifically, doesn't do any video/picture analysis. Again, this is something we can already do -- Google Goggles and other image/video analysis programs and algorithms. Could it? Sure, at a massive increase in cost, but we already know this can be done, so it didn't need to be added on to this proof of concept.

But Watson can analyze data (in this case, language), relate it to knowledge already within its database, make determinations about the context of the variable/word (run like a foot race, or run for office? Fried like an egg, or fried like a sunburn?), and not only come up with correct output, but assign a probability to it based on the knowledge it has.

Watson is a computer making an educated guess, not simply searching for a matching answer to a given problem like a calculator.

Now, how does that have real world applications?

In the short term, it's a new form of search and data organization. Doesn't seem like a big deal, but that's what Yahoo thought when a startup called Google had a new way of searching and organizing data.

In the long term? Far reaching applications. Take, for example, home care of dementia/Alzheimer's patients. The poor oldster has turned on the sink, wet their hands, put some soap on them...and has no idea what they were doing. And yes, I've seen that happen.

Watson v8.0 with contextual video and audio processing and an array of sensors in the home sees/hears/detects all of these clues, and maybe the ones before it like a toilet flush 35 seconds prior, and that the oldster entered the bathroom and headed for the toilet, and that the sink is in the bathroom, so water isn't being run to wash dishes or thaw food, and that it was cleaned recently, so that's not why the water is running, etc, etc...

...and tells the oldster "You just went to the bathroom, and you are washing your hands. Rinse your hands in the running water, and turn off the water by turning the handle."

Or it's a miniature version that can be trained to detect brainwave patterns in ALS patients (or even coma patients), and output speech. Or you can just speak into it, if you can do that. And the output is in whatever language the user wants it to be in, because it can not only analyze the brainwaves, but it can recognize slang, turns of phrase, and what have you.

That is what Watson is the beginning of. Right now, it's 100 high powered servers that cost more than most of us will ever see in a lifetime. Given Moore's Law and the exponential growth of computing power, it isn't too long before $25 million Watson is a $200 iThink.

This is a proof of concept of a large breakthrough in data analysis. It is at the pinnacle of achievement, and is an exquisitely crafted machine designed by some of the greatest minds of our time. I look at it in awe like I would a Michelangelo, or Cristo Redentor, or when I hear one of the great classical compositions.

It's not something to replace humans, but is a wonderful, beautiful creation of them to enhance the experience of humanity. Jeopardy is just the beginning.

Originally posted to Explodingkitchen on Tue Feb 15, 2011 at 10:11 AM PST.

Also republished by SciTech and Community Spotlight.

Your Email has been sent.
You must add at least one tag to this diary before publishing it.

Add keywords that describe this diary. Separate multiple keywords with commas.
Tagging tips - Search For Tags - Browse For Tags


More Tagging tips:

A tag is a way to search for this diary. If someone is searching for "Barack Obama," is this a diary they'd be trying to find?

Use a person's full name, without any title. Senator Obama may become President Obama, and Michelle Obama might run for office.

If your diary covers an election or elected official, use election tags, which are generally the state abbreviation followed by the office. CA-01 is the first district House seat. CA-Sen covers both senate races. NY-GOV covers the New York governor's race.

Tags do not compound: that is, "education reform" is a completely different tag from "education". A tag like "reform" alone is probably not meaningful.

Consider if one or more of these tags fits your diary: Civil Rights, Community, Congress, Culture, Economy, Education, Elections, Energy, Environment, Health Care, International, Labor, Law, Media, Meta, National Security, Science, Transportation, or White House. If your diary is specific to a state, consider adding the state (California, Texas, etc). Keep in mind, though, that there are many wonderful and important diaries that don't fit in any of these tags. Don't worry if yours doesn't.

You can add a private note to this diary when hotlisting it:
Are you sure you want to remove this diary from your hotlist?
Are you sure you want to remove your recommendation? You can only recommend a diary once, so you will not be able to re-recommend it afterwards.
Rescue this diary, and add a note:
Are you sure you want to remove this diary from Rescue?
Choose where to republish this diary. The diary will be added to the queue for that group. Publish it from the queue to make it appear.

You must be a member of a group to use this feature.

Add a quick update to your diary without changing the diary itself:
Are you sure you want to remove this diary?
(The diary will be removed from the site and returned to your drafts for further editing.)
(The diary will be removed.)
Are you sure you want to save these changes to the published diary?

Comment Preferences

  •  Work Watson down to a managable size... (4+ / 0-)
    Recommended by:
    caul, rb608, khereva, palantir

    ...or team it with a reasonably-sized mobile device, and it might work. But it may be years before a Watson-based device would be available for helping someone with dementia, and you could pay for a small army of human domestic help with what has built Watson.

    Float like a manhole cover, sting like a sash weight! Clean Coal Is A Clinker!

    by JeffW on Tue Feb 15, 2011 at 10:17:41 AM PST

    •  The smartphones we have today... (8+ / 0-)

      ...are more powerful than PCs 10 years ago, and they have more computing power than some of the older mainframe systems.

      Moore's Law and the exponential growth of computing power doesn't take long to miniaturize things.  A $200 device might be pushing it, but as a few thousand dollar system wired into a personal or nursing home?  Not that far out of reach, and I mean less than 20 years, if not sooner.

      And that's just one application where being able to process contextual clues would come in handy, and may not even be the best example.

      •  Power does not equal intelligence... (3+ / 0-)
        Recommended by:
        caul, palantir, 207wickedgood

        ...otherwise all the Crays in research would be providing rough drafts of the papers from the data they are crunching. This is a lot like extrapolating the stuff they said we'd see by now: Moon colonies, jetpacks, bubble-top flying cars that run on smal fusion reators, etc.

        I used to subscribe to a magazine called AI Journal, because I was hoping to employ AI algorithms for "intelligent" traffic control, with regard to computerized traffic signal systems. The magazine folded after a couple of years, as the promise just wasn't there yet. And I never did get to do anything with the computer systems we installed for our traffic signals, but that was a failure of human intelligence!

        To me, this was like the 1950's demo of  UNIVAC I predicting the winner of the Presidential election. A tool being used as a toy. Color me only slightly impressed.

        Float like a manhole cover, sting like a sash weight! Clean Coal Is A Clinker!

        by JeffW on Tue Feb 15, 2011 at 11:23:38 AM PST

        [ Parent ]

      •  Moore's Law won't last forever... (8+ / 0-)

        Currently, some of the characteristics of integrated circuit fabrication are measured not in microns, but in atoms (as in "most 24nm technologies use a substrate that's 40 atoms thick").   Clearly, we're nearing the end of the road.  Reducing the size of Watson to a deck of cards is not something I'd plan on.

        There's something else that bothers me about all this... great minds with a large budget have been able to conquer two of the great problems facing humanity as we enter the 21st century: chess and Jeopardy.  What do you think it will take to do something significant?

        Don't be a DON'T-DO... Be a DO-DO!

        by godwhataklutz on Tue Feb 15, 2011 at 03:15:02 PM PST

        [ Parent ]

        •  They said that a while ago... (7+ / 0-)

          ...when processors starting breaking the 3ghz mark.  That we wouldn't continue to keep doubling computing power.

          Well, we did.  Just not in the way expected.  Instead of having a single CPU running at 3ghz that can handle a single process at a time, you have 4 CPUs on one chip running at 3ghz, each able to handle two (or more, depending on architecture) processes at once rather than one.

          You're also looking at the end goal of Deep Blue and Watson of beating a grandmaster and competing on Jeopardy.  The promise behind them isn't the showing off that's done to drum up interest in the project.

          Deep Blue had to handle a game tree set at ~10^123 possible games and solve for Shannon's Number.  Even then, it didn't have the horsepower to crunch all of it in a reasonable amount of time to "see every game" -- but it saw enough moves ahead to defeat a grandmaster.

          Watson can understand context, and can realize that cash is a term for money, an action you can do (cash a check), slang for cool or awesome, etc... on the fly by analyzing how the term is used, rather than each definition having to be manually linked and defined.  That's a huge breakthrough, and while the sort of processing power behind it seems like it's a lot, it's relatively not.  10 cabinets of rackmounted servers is a miniscule amount of space for the amount of power that's being pumped out.

          Being able to play Jeopardy provided focus on the project, and serves as a useful side effect of what proper contextual data analysis can do that drums up interest in the research.  The significance comes from seeing the abilities of the machine, then figuring out how to apply it to another use.

          •  Understanding context? (1+ / 0-)
            Recommended by:
            Pris from LA

            The ability for software to distinguish between noun and verb forms of "cash" was prototyped thirty years ago.  That said, natural language processing is (very very) hard... most success stories rely on a focused universe (e.g. newspaper stories about car accidents)... the Watson demonstration is the next step on this journey
            The future you envision is a long way away.  We're going to need something beyond CMOS, and we're going to need to take a couple more steps (at 20 years per) to get to where an automaton can deal with natural (spoken) language in general.
            Watson is an interesting project and I hope IBM doesn't lay off the researchers now that it's done.

            Don't be a DON'T-DO... Be a DO-DO!

            by godwhataklutz on Wed Feb 16, 2011 at 05:35:13 AM PST

            [ Parent ]

      •  as I was watching it tonight with my (2+ / 0-)
        Recommended by:
        Angie in WA State, Pris from LA

        14 year old, I commented to him that in 20 years, he'll be holding Watson in his hand, just like 20 years ago, the future was what my smartphone looks like now...

        Whenever the people are well-informed, they can be trusted with their own government. ~Thomas Jefferson

        by k8dd8d on Tue Feb 15, 2011 at 09:10:07 PM PST

        [ Parent ]

      •  Why keep it in hardware? (4+ / 0-)

        Just make it a web service like Google and you can run it on today's devices.

        Change comes fast and change comes slow. But change come. --Caroline, or Change

        by Mangala on Tue Feb 15, 2011 at 09:13:11 PM PST

        [ Parent ]

        •  It's still too hefty for that. (1+ / 0-)
          Recommended by:
          Pris from LA

          That's a market price of ~$25 million of hardware to handle a single query at a time in a short amount of time.

          And that's only to access its current functionality, rather than any new applications.  It might be worth that much for the medical diagnostic/treatment applications, but that would be a serious expense with a hefty pricetag per query.

          •  Won't always be that hefty. (1+ / 0-)
            Recommended by:
            Angie in WA State

            To wit: $200 netbooks that are more powerful than $3,000 ultra-compact notebook computers 10 years ago. I have two specimens which are illustrative of this new reality: an IBM Thinkpad 240, and an Acer AspireOne Atom-based netbook. They are close to the same size and weight, but one wipes the floor with the other. I tried to keep the 240 useful with a light Linux load, but the Acer is just so much better. And Ubuntu runs like a bat out of hell on it. It is still a netbook. It's not a full-fledged notebook computer. It has a single core and the Atom chip. has lots of compromises made for battery life, energy consumption and cool running for the netbook platform. However: $200 vs. $3,000. In ten years.

            Health care should not be a privilege for the few, but a human right for all.

            by Pris from LA on Wed Feb 16, 2011 at 12:56:08 PM PST

            [ Parent ]

  •  Thank you for a good diary (9+ / 0-)

    I find Watson very interesting, and you helped explain more of it.

    Join us at Bookflurries: Bookchat on Wednesday nights 8:00 PM EST

    by cfk on Tue Feb 15, 2011 at 10:32:38 AM PST

  •  thanks for the diary and especially... (5+ / 0-)

    ...for being clear on what is and is not intelligence.

    "I hope to have God on my side, but I must have Kentucky." -Abraham Lincoln

    by jethropalerobber on Tue Feb 15, 2011 at 10:54:40 AM PST

  •  Watson would've been a lot more entertaining (10+ / 0-)

    if it put on its best Sean Connery impression, asked Alex Trebek about penis mightiers and insulted his mother.

    "Not in the Rs? That's not what your mother said last night Trebek!"

    Ahhh....the days when SNL was funny.

    Twitter: Dumbing down American politics, 140 characters at a time

    by yg17 on Tue Feb 15, 2011 at 12:02:29 PM PST

  •  NOVA did a great special on this last week (7+ / 0-)

    NOVA did an interesting special on Watson last week.  I can't find a link to the show, just  this page on NOVA.  A remarkable machine, and NOVA did a good job explaining why Watson was able to do what he did, and the challenges they faced (and also many dead-ends that other AI projects have tried).

    Minority rights should never be subject to majority vote.

    by lostboyjim on Tue Feb 15, 2011 at 02:19:52 PM PST

  •  I understand that the purpose for Watson (11+ / 0-)

    that they are projecting is more akin to medical diagnosis as in House. Because no doctor can keep up with the medical literature, a Watson-like device would search for the symptoms and combination of symptoms  and give the doc a heads-up about some possible diagnoses.

    The decision would still be up to the doctor, but at least he/she would have access to more of the pertinent information.

    Conservation! Because the cheapest energy is the energy you don't use.

    by ohiolibrarian on Tue Feb 15, 2011 at 02:26:56 PM PST

    •  Yep. (6+ / 0-)

      Differential diagnosis is evaluating symptoms (a data set) in context, and deciding what it can't be in the hopes you narrow it down to what it can be.

      And you've hit it on the nose -- it still works with a human. Watson and other specialized "intelligence" machines aren't there to replace humans.  They are there to complement their abilities.

      The other big key is being able to give an accurate idea of how confident the machine is in its answer.  That's more a breakthrough than you think.  After all, if the computer spits out 36 possible answers, that's not that useful.

      If it spits out 36 answers and is able to put an accurate probability of correctness on them, it makes it far more useful.

  •  Watched last night, enjoyed it. (6+ / 0-)

    But I absolutely agree that people talk about it in the context of Jeopardy. We ignore the fact that researchers has created a computer that can retrieve accurate information in seconds. The impact it could have on our lifes has incredible potential, and we can't ignore that for the sake of a game show.

    Here's hoping IBM continues to refine this device and make it more useful for the rest of us.

  •  wow (1+ / 0-)
    Recommended by:

    how awesome would this be?

    .and tells the oldster "You just went to the bathroom, and you are washing your hands. Rinse your hands in the running water, and turn off the water by turning the handle

    Thanks for the intriguing diary

    When will our consciences grow so tender that we will act to prevent human misery rather than avenge it? Eleanor Roosevelt

    by IndyRobin on Tue Feb 15, 2011 at 02:35:43 PM PST

  •  I don't follow the "massive" increase... (2+ / 0-)
    Recommended by:
    Explodingkitchen, caul

    in cost for adding an up-front audio processor. I can do that for less than 100 bucks with a single microprocessor. Everything else is downstream.

    Perhaps the rationale has more to do with (as you also suggested) that it adds no novelty to Watson.

    The man who moves a mountain begins by moving away small stones. -Confucius

    by Malachite on Tue Feb 15, 2011 at 03:08:09 PM PST

    •  It would also take Watson a bit longer (1+ / 0-)
      Recommended by:

      to process the sound, and thus take away his ability to buzz in before the other players... which seemed impressive until/unless one thought about it.

      sig deleted before the great DK4 deluge. you all know what it says anyway.

      by khereva on Tue Feb 15, 2011 at 03:37:36 PM PST

      [ Parent ]

      •  He doesn't buzz in before he others... (1+ / 0-)
        Recommended by:
        Jeff in CA

        ...unless he's calculated the answer.  During playtesting, humans would out buzz him on clues that did not take a long time for the host to read.  Watson's processing isn't real time yet, and the confidence level needed is an adjustment for him buzzing in.

        If it was truly an issue, he'd be dominating the other two rather than them getting a lot of the buzzes in.  It'd be like saying Ken Jennings had an unfair advantage against other humans because he combined hellacious buzzer reflexes with the ability to make an educated guess on stuff he didn't even know about.

    •  Because audio-processing still sucks... (1+ / 0-)
      Recommended by:

      Google can't properly translate my voicemail to text, and they've got a ton of horsepower behind the software that does that.  You'd have to throw a massive amount of hardware behind it so that it could process more surrounding information, and it massively increases the amount of data to consider.

      On top of that, it would more specialize Watson towards playing Jeopardy rather than an all around contextual data analysis system.  Sure, if IBM spent the money, Watson could have realized that Ken Jennings answered the 20s incorrectly before it did it.  But that is extra hardware, processing algorithms (processing the input, understanding that that input at that time is an answer, understanding that that answer is incorrect, re-evaluating the data with that new information which resets the crunching clock) that doesn't affect its main purposes that it's showing off -- contextual data analysis.

      Playing Jeopardy is a side-effect of Watson's algorithms, not the true purpose of them.  Jeopardy provides the limit on how fast Watson needs to generate an answer, giving the desginers a goal to shoot for.

      •  Watson's algorithms (3+ / 0-)

        My understanding is Watson doesn't have "an" algorithm.  Instead, many different algorithms are used in parallel and they vote for a final answer.  That in itself is an algorithm, of course, but not exactly in the normal sense.

        •  In a rough sense... (2+ / 0-)
          Recommended by:
          Jeff in CA, Pris from LA

          ..that's a close description.  A lot of Watson's ability has to do with the horsepower behind it.  High speed disks and/or fiber-channel storage, the ability to process thousands of tasks at once rather than the 2 to 4 a consumer grade computer does, etc.

          It's what's known as an embarrassingly parallel algorithm(s)/program(s).  They can very easily split it into numerous parallel tasks.  How they pull it all together with a master algorithm or the like is well above my pay grade and knowledge, however.

          •  Rough sense (0+ / 0-)
            rather than the 2 to 4 a consumer grade computer does, etc.

            You are talking how many cores commodity hardware has. In fact multi threaded apps greatly extend "parallelism". Example: While one thread is twiddling its thumbs waiting for the painfully slow disk to finally seek to the requested data, the thread sleeps while other threads work.

            How they pull it all together with a master algorithm or the like is well above my pay grade and knowledge, however.

            This will get you started:

  •  Yes, Watson is great at playing Jeopardy, (6+ / 0-)

    but it will never be able to write a sequel to the song parody I Lost on Jeopardy that's nearly as funny as the one I'm hopeful Weird Al will write.

    Barack Obama in the Oval Office: There's a black man who knows his place.

    by Greasy Grant on Tue Feb 15, 2011 at 03:16:28 PM PST

    •  Why would you want it to? (4+ / 0-)

      Surprisingly, given the proper dataset, computing power, and algorithms, it probably could.  We already have a formula for how to write a radio hit (intro,verse,chorus,verse,chorus,bridge/breakdown,chorus,chorus,out between 3:15 and 3:45 long, pick a popular genre).  We know what makes things funny, although that taste obviously differs between people.  In fact, it could probably analyze Weird Al's specific humor, his past songs, etc...and come reasonably close.

      That's all a massive data set.  But there's not a practical use for it.  I think that people get the wrong idea about computing intelligence -- it's not meant to be creative, or replace human thought.  It's there to augment it.

      With the above project...all you end up getting is a highly specialized Weird Al digital impersonator.  It's useless, except as a proof of concept of contextual analysis -- which we already have here.  It's just extrapolating it to a larger data set (words, lyrics, song structure, contextual language) rather than just contextual language.

      •  Ummm... I was being somewhat tongue-in-cheek. n/t (3+ / 0-)

        Barack Obama in the Oval Office: There's a black man who knows his place.

        by Greasy Grant on Tue Feb 15, 2011 at 03:59:52 PM PST

        [ Parent ]

      •  Excellent overall (3+ / 0-)
        Recommended by:
        Odysseus, caul, shenderson

        but I have to pick a bone with you on one point:

        I think that people get the wrong idea about computing intelligence -- it's not meant to be creative, or replace human thought.  It's there to augment it.

        At the risk of sounding like a Luddite, I would say it is there to save someone the cost of paying someone else a paycheck. So whether it is creativity or human thought or simple algorithmic processing of huge data sets (and without departing into the philosophical brambles concerning whether they are in fact three different things to begin with) doesn't really matter, it is intended to replace people. Decades ago, when we thought that would mean more leisure for everyone, it could be assumed it would be a good thing. Now that we know it will only mean more money for people who already have too much "leisure" and less leisure for people who will never have too much money, I'm not so sure.
        •  I won't argue... (1+ / 0-)
          Recommended by:

          ...with you on how they end up being used, because it is something I really can't disagree with.

          What I was more implying was that we likely won't end up with a true "intelligent" computer that acts like a human a la Data from Star Trek.  We'll create one, I'm sure, but more to say we did it than to have a practical purpose for them.

          We're better of creating specialized intelligences for specific tasks, or to coordinate a massive amount of tasks, rather than trying to make a mechanical/electronic human.

          Efficiency as a word means less work or less effort needed, and yes, that does mean that humans do get replaced in some tasks.   After all, if you have a network of systems that can manage not just the airspace and takeoff/landing patterns at one airport, but a network that manages the entire airspace of the US and/or world, you'll probably need fewer air traffic controllers, and gain a few systems administrators.

          I won't lie, when I did consulting for a living, I automated a few junior systems administrators out of jobs.  It's part of the reason why I've settled into a salaried position where my job isn't to put other people out of work.

          The smart plan isn't to prevent it from happening, but to plan where the displaced workers will land in the economy, how to ease the landing, and how to benefit the most people from the transition.  I'm a liberal because I want to see the government plan that and provide help, rather than tell those displaced to learn something new and leave them to starve.

    •  I'm just hoping Watson can replace glenn beck (3+ / 0-)
      Recommended by:
      caul, Greasy Grant, 207wickedgood

      It'd certainly be more sane and perhaps more entertaining as well.

      A learning experience is one of those things that says, 'You know that thing you just did? Don't do that.' Douglas Adams

      by dougymi on Tue Feb 15, 2011 at 04:19:51 PM PST

      [ Parent ]

    •  And it won't blog (1+ / 0-)
      Recommended by:

      about it's Jeopardy experience.

      Artists to my mind are the real architects of change, and not the political legislators who implement change after the fact. -William S. Burroughs

      by hotdamn on Tue Feb 15, 2011 at 05:19:25 PM PST

      [ Parent ]

  •  Until it can figure out how to elect (5+ / 0-)

    Democrats and defeat Republicans, I won't be impressed.

  •  Seems to me that Watson is the first (3+ / 0-)
    Recommended by:
    Explodingkitchen, palantir, Mangala

    building block needed to build something like the Ship Computer on Star Trek.  

    •  We can already build the computer from Star Trek.. (5+ / 0-)

      ...we just can't build all the systems for it to control like a warp drive.  And the input would probably look like greek to anyone who doesn't program or script.

      At my current gig, part of my job is working with clusters of servers that all respond to one master that drives them.  I issue a set of commands to the master server, and it takes those commands, interprets them, and passes them on to the proper clusters, which then do the real work, and send an output back to the master, which reports back to me in a nice, concise email.

      It's much easier to split up computer tasks into groups of specialized servers, then have a bastion/master server that pulls the puppet strings, so to speak. The computer in Star Trek, I imagine, would be more of a bastion/master rather than a monolithic machine that controlled everything.

      Because the last thing you'd want is a buggy holodeck program to crash the entire ship, rather than just the holodeck display cluster.

  •  A political comment... (2+ / 0-)
    Recommended by:
    Explodingkitchen, palantir

    The fact that Watson is an output of IBM is actually incredibly rare. Almost ALL basic research that turns into applied technology is funded by the government and developed in government labs or heavily subsidized universities.

    I am considering a project explaining this that would use crowd sourced research to help me make this crucial point clear.

    •  Watson isn't solely an IBM project... (3+ / 0-)
      Recommended by:
      Gary in NY, caul, palantir

      They also worked with some leading minds at leading public and private universities.  It was financially backed by IBM, however.

      Computing is one of the rare areas where private discovery tends to happen at the same rate as public discovery.  One, IBM and other computing firms that do research make enough money that they can afford to financially back such a project and expect to come up with something to make a profit off of (and therefore can take the risk), and a lot of the great computer science minds don't stay in academia and research there, and instead move on to the private sector because the pay differential is so great (since computing brings in such profit).

      To give you an idea of the pay differential between the two sectors when it comes to a computer related job, I have less than a decade of experience administrating enterprise systems, and an industry certification.  I don't have a college degree.  I don't create new systems, I just take existing products and platforms and turn them into something the people who actually create applications can use.

      I make more than the average computer science professor that has a PhD, a decade of post-doctoral research, and I didn't have to fight for a limited amount of tenure-track positions in an environment where most universities are having their funding cut, and I guarantee I work less hours in a week than they do.

      I'm not saying that to brag, but just to point out that someone going into it for academia has to really like academia, because the knowledge that you have to have to do research at that level pays two and a half to four times in the private sector than it does in academia.

      Because of that pay difference, a lot more computing research comes out of private business than most other fields.  Then again, they're paying a lot more for the research.

      •  Fair enough, but it's the exception (4+ / 0-)
        Recommended by:
        Odysseus, Explodingkitchen, caul, palantir

        that proves the rule.

        Nanotech, molecular biology, alternative fuels... you name it - where there's any question of ROI, private firms simply will not engage in research. They'll just free-ride on government research when it nears market stage.

        •  Oh, I agree completely with... (2+ / 0-)
          Recommended by:
          caul, palantir

          ...the political sentiment in your post.  I was just pointing out how the computing industry and research within it differs a lot from the normal research rules.

          There's other oddities about the industry as well, like how little a college degree means at a lot places.  Has a lot to do with how much of the modern industry was built by college dropouts.  I work next to guys that graduated from MIT, Stanford, and other prestigious schools, and I hold my own just fine.   They have a wider base of knowledge, so they know things I don't, and I've been hands on longer, so in a lot of subjects, I have a greater depth of knowledge, so I'll know things they won't.  But when it comes down to it, outside of a few companies like Google that want to promote education and therefore care about your GPA after you graduated 20 years ago, most IT shops simply treat a bachelors as the equivalent of 3 to 4 years of experience.

  •  WATSON is The Computer in (4+ / 0-)

    Deskset  Impossible with 1950's technology built for real

    I'm the terror that blogs in the Night,. and the daytime too.

    by JML9999 on Tue Feb 15, 2011 at 05:26:09 PM PST

  •  Great (2+ / 0-)
    Recommended by:
    caul, palantir

    more jobs lost to technology.

    I watched the NOVA documentary on Watson.

    It is very good at going through facts at great speed and then shifting through the results and coming up with the most probable answer.

    Can Watson tell us why humans are so insane they are intent on killing themselves off and destroying the planet?

  •  Something I'm not sure was recognized by most (5+ / 0-)

    watching Jeopardy wrt Watson.

    ALL the contestants on the show knew the answers to just about ALL the questions.  What allowed Watson to dominate is that the hold off of the contestants' signaling devices is something even the simplest computing device could master better than the best human.  Its reaction time is measured in small fractions of a millisecond, whereas the real life humans have no such ability.  For that reason, I think the show is a stunt.  

    In other words, the answers that I saw on tonight's show at least were too easy - I knew upwards of 80% of them, and I'm pretty sure I was not the most knowledgeable of the collection of human contestants.  However, it was pretty evident that Watson had a clear advantage in the timing aspect of buzzing in that was insurmountable.  

    That said, I would agree that the data mining potential of computers is large.  However, I don't see this as a potentially ubiquitous device.  I think of it as having a much more directly profit making application - a means to manipulate or profit from the stock market for example, or voters.  

    No conspiracies here, just a much more direct connection between the primary motivations of those who can afford to harness this technology in the near term.

    Certainly from our standpoint, this gives us a sense of momentum -- when the United States has accolades tossed its way, rather than shoes. - PJ Crowley

    by nsfbr on Tue Feb 15, 2011 at 06:38:03 PM PST

    •  This was probablly the central struggle (7+ / 0-)

      to making Watson vs the experts compelling television. Although natural language was the obstacle to computer competing on Jeopardy!, once that was conquered, they have an amazing advantage in speed.

      You could see it in certain categories that Watson dominated and did poorly. He was able to recall all of the Beatles songs, but struggled MIGHTILY with the double-meaning words (confusing apex with peak, chic with class, etc.)

      80% would be pretty good, almost amazing. You should go on the show, if what you say is so.

      Where Watson truly dominated was the answers that was just beyond human remembrance. What you are thinking of is the questions that were like, "Hanson's disease is also known as this?" Watson quickly re-called Leprosy, while most humans (yourself included) were like, oh, I knew that! When in fact you had heard it once and discarded/forgotten it.

      Watson did much better on the second round, strictly because the questions were harder. The questions were on things that most humans don't know (yes, even the uber-geniusii of Rutter and Jennings) or things that take a long time to recall.

      •  The 80% (4+ / 0-)

        My metric is if I can annoy my wife by saying the answer out loud before the correct answer is spoken (or tonight by that annoying three choice thing from Watson being shown on screen).  I could never go on the show due to the angst of running into one of "those" shows where the categories are just completely outside of anything I'd know - I hate that.

        I'm also pretty sure that sitting on my couch, a glass of red wine at my side I am as relaxed as I can be.  Put me on stage, hmm, 40%?  50%?  I have no idea.  It is a much bigger deal to do it in front of an audience.  At least for me.

        Certainly from our standpoint, this gives us a sense of momentum -- when the United States has accolades tossed its way, rather than shoes. - PJ Crowley

        by nsfbr on Tue Feb 15, 2011 at 07:08:49 PM PST

        [ Parent ]

    •  I agree, it looked like they knew (1+ / 0-)
      Recommended by:

      the answers many times, just came in later signaling

      "I've taken up sculpting recently. Landscapes mostly." ~ Yogi Bear

      by eXtina on Tue Feb 15, 2011 at 08:37:25 PM PST

      [ Parent ]

    •  Watson's overall reaction time was slowed down (4+ / 0-)
      What allowed Watson to dominate is that the hold off of the contestants' signaling devices is something even the simplest computing device could master better than the best human.  Its reaction time is measured in small fractions of a millisecond, whereas the real life humans have no such ability. was pretty evident that Watson had a clear advantage in the timing aspect of buzzing in that was insurmountable.

      Because Watson had to depress a buzzer just as the 2 humans did, it was visibly clear that the overall reaction time of Watson was greatly increased above mere milliseconds.  To me this seemed an intentional measure to make it seem more "fair" to compete against humans.  

      Think how much faster Watson's reaction time would have been if it simply sent electrons to the buzzer at a fifth of the speed of light.

      Never before in all our history have these forces been so united against one candidate... They are unanimous in their hate for me. And I welcome their hatred.-FDR

      by Jeff in CA on Tue Feb 15, 2011 at 08:58:53 PM PST

      [ Parent ]

  •  IBM's next grand challenge (7+ / 0-)

    Interpreting Sarah Palin - or learning to understand un-natural language.

    2010: An Unforced Error Odyssey

    by Minerva on Tue Feb 15, 2011 at 06:43:21 PM PST

  •  I wonder how Watson (5+ / 0-)

    fucked up the last question this evening. Seemed pretty simple.

  •  Would this have a practical application (3+ / 0-)
    Recommended by:
    Explodingkitchen, caul, Pris from LA

    in the diagnosis of illness?

    What if doctors all over the country fed it all the data they had on a particular patient who could have any number of conditions?

    Could it be programmed with all the medical knowledge available which it compares to the data received and provide a diagnosis with probability of correctness?

    Each time a doctor fed in data about the same condition it would be added to the database and analyzed against all the data received concerning patients with similar symptoms.

    Does this sound like a realistic application?

    The community of fools might be small if it were not such an accomplished proselytizer.

    by ZedMont on Tue Feb 15, 2011 at 08:10:19 PM PST

    •  Yes. (2+ / 0-)
      Recommended by:
      Pris from LA, ZedMont

      In fact, it's the application that IBM is most excited about.  Watson can analyze reams of data faster than a full team of diagnosticians.

      But just because we have the power to make something like this, doesn't mean we can.  We can't just have it analyze -all- the data in the record, it needs to be able to respect privacy laws of various jurisdictions, so instead of just going full bore at the problem, we also have to find a balance with the right to privacy.

      Breaking it up into separate data sets that stop the systems and your personal characteristics from being tied to you having the disease (at least without you signing consent to have it shared) will prove to be a difficult challenge, and will add to the complexity of the problem.

      •  I don't understand why doctors could not (1+ / 0-)
        Recommended by:
        Pris from LA

        just submit a pseudonym (Patient X, whatever).  The doc is only interested in getting back diagnostic data, because he will still be the one making the official diagnosis - unless of course, this thing is so damn smart it goes to medical school and passes the boards.

        The machine itself doesn't need to know the real name of any patient to analyze the data and transmit its findings, does it?

        The community of fools might be small if it were not such an accomplished proselytizer.

        by ZedMont on Wed Feb 16, 2011 at 07:17:53 AM PST

        [ Parent ]

  •  As a Jeopardy Geek... (2+ / 0-)
    Recommended by:
    Explodingkitchen, Pris from LA

    And a tech enthusiast, I've been following this for months. I'm a bit disappointed that they haven't yet thrown one of the trickier categories at Watson yet (like Before and After) to test his language prowess (though he did pretty well at some of the gnarlier literary questions.)

    As far as question difficulty goes, it's been about median level. Mind you, this fits into the Jeopardy paradigm pretty well. Raise the stakes as the game goes on, whilst keeping the audience at home somewhat interested by throwing out a few questions that they might know. Be a vicarious Jennings.

    With regards to the possibilities, the creative classes (engineers, lawyers, etc.) could see huge benefits. Imagine a responsive and socially contextualized database. It'd certainly lead to an increase in efficiency.

    So yes. Software that allows a computer to become a genuine social actor and tool that would augment (not necessarily compete with) human intelligence greatly intrigues me. Watson does have many shortcomings, but if Moore's Law holds (and yes, we're reaching its physical limits, but computing has a way of innovating at a ludicrous rate), we may wind up with a much smarter smartphone (which incidentally sounds like a Jeopardy category.)

  •  Yeah, but he/it thought Toronto (3+ / 0-)
    Recommended by:
    Explodingkitchen, caul, Pris from LA

    was a US city?!? How is that possible?

    "I've taken up sculpting recently. Landscapes mostly." ~ Yogi Bear

    by eXtina on Tue Feb 15, 2011 at 08:36:20 PM PST

    •  If his decision making isn't weighted... (2+ / 0-)
      Recommended by:
      HudsonValleyMark, Pris from LA

      ...enough towards taking the category into context, then that could be why.

      I don't remember them showing the probabilities on the Final Jeopardy answer, but apparently the response indicated that Watson didn't meet his confidence level.

      •  yeah, the ????? and amount bet (1+ / 0-)
        Recommended by:
        Pris from LA

        I don't know what algorithm Watson was using to choose a dollar figure there, but probably something to do with expected value, bearing in mind the strategic context (which, of course, was that Watson had a crushing lead). To bet less than $1000 probably indicated that Watson was nowhere near the "buzz threshold."

        Of course I can't rule out a "mercy rule." :)

  •  And so it begins... (2+ / 0-)
    Recommended by:
    Explodingkitchen, palantir

    now I know what that guy from Maryland was going on about when he said if we gays were allowed to marry, next people would be marrying their androids.  I bet somebody already has dibs on Watson.

    Seriously, very interesting diary.  I honestly think that given the exponential increase of knowledge versus the typical lifespan of a human and a human's typical ability to remember/recall facts, we are going to need computers like Watson, that can aggregate knowledge and never forget anything, to be able to make the jump to the next level.

  •  Great Diary, and obviously the power of (3+ / 0-)

    Watson and it's future applications is amazing.

    But as far as Jeopardy is concerned, the real decision of who will win was made when the questions and categories were selected.  The NOVA program made it clear that when the question is factual based, and doesn't require much processing (as in who was King of England in 1546 from tonight's show) it is always going to win because it can process so much faster.   When the question is more convoluted, or uses unusual phrasing, etc, then the humans will be faster (as in missing the Harry Potter question last night...the reference was to Voldemort, but Voldemort's name is almost never used in the books, so coming up with the link of who killed three characters was not something it could solve with certainty).

    So how were the categories and questions selected?  Did IBM have any input to that process?  Did they follow an algorithm for how Jeopardy is normally written, etc?  I have not seen that discussed anywhere.

    Whenever the people are well-informed, they can be trusted with their own government. ~Thomas Jefferson

    by k8dd8d on Tue Feb 15, 2011 at 09:20:55 PM PST

  •  Nice diary, thank you! (3+ / 0-)

    Computers are my avocation, so this article is greatly appreciated.

  •  Would you mind if I modified your tags (2+ / 0-)
    Recommended by:
    Explodingkitchen, Pris from LA

    I'd add Technology and Science

  •  How does this relate to the path Kurzweil laid (1+ / 0-)
    Recommended by:
    Pris from LA

    out in his book, "The Singularity is Near?"

    NOVA was saying Watson may be the precursor to complex medical diagnostics as a doctor's aid (and who knows what else).

    It sure fuels the imagination.

    "Dega dega dega dega. Break up the concrete..." The Pretenders

    by Terra Mystica on Wed Feb 16, 2011 at 04:31:05 AM PST

  •  Bittersweet accomplishment (1+ / 0-)
    Recommended by:
    Pris from LA

    I enjoyed watching Watson play Jeopardy, but the experience takes on a different meaning when you realize that this may be the last big accomplishment for IBM in the US, and that naming the computer Watson may be the only remaining link to that name in a few years.

    IBM is indicating that the future of their research development will not be in the US, and has followed 2 years of massive offshoring with the start of their centennial celebrations -  in Bangalore, with the main event rumored to follow -  in Dubai.

    It's likely that all the highly anticipated real-world applications for this exciting technology will be coming from China, India, or whichever country has the cheapest labor force at the time.

    I'd be curious to know what Watson thinks the results of this business strategy might be.

  •  Here's why I like Watson (1+ / 0-)
    Recommended by:
    Pris from LA

    I can envision Watson as the final arbitrator on talk radio.

    If a rightwing caller wants to engage me, the host, in a typical reprehensible liefest, an instantaneous consultation with Watson will quickly deliver the real facts.

    Which always seem to have a liberal bias.

    Fuck with the truth at your own peril. - Anonymous

    by thenekkidtruth on Wed Feb 16, 2011 at 09:07:38 AM PST

Subscribe or Donate to support Daily Kos.

Click here for the mobile view of the site