Forgot your password?
typodupeerror
AI Robotics Science

Teaching Robot Learners To Ask Good Questions 93

Posted by Soulskill
from the why-must-you-kill-all-humans dept.
garthsundem writes "I disagree with this article's opening line: 'Within a decade, personal robots could become as common in U.S. homes as any other major appliance.' Haven't we been promised this since the 50s? But I'm fascinated by the rest — how do you teach humans to teach robots? Or, more precisely, how can you teach robots to teach humans to teach robots? The idea that designers can put a flexible platform in a robot, allowing users to determine functionality, is pretty interesting. The lead researcher for this project said, 'People are not so good at teaching robots because they don't understand the robots' learning mechanism. It's like when you try to train a dog, and it's difficult because dogs do not learn like humans do. We wanted to find out the best kinds of questions a robot could ask to make the human-robot relationship as 'human' as it can be.'"
This discussion has been archived. No new comments can be posted.

Teaching Robot Learners To Ask Good Questions

Comments Filter:
  • I think the only thing keeping them out of homes NOW is no used market yet, and the high cost ($200-$5000) of entry. Basically the same problem that faced real personal computers in 1979.

    Only THAT will change the real adoption of robots into the average American household.

    • by ae1294 (1547521) on Friday March 09, 2012 @02:52PM (#39303553) Journal

      No what's keeping them out of homes right now is their unquenchable thirst for human blood...

      • by geekoid (135745)

        Hey! put that routine in as a joke, don't blame me if QA didn't do their job.

      • by nman64 (912054) *

        Perhaps if we throw wave after wave of our own men at them, they'll reach their kill limit and become docile workers...

    • It's more than price - it's functionality. At the moment you can get human help for x dollars that can clean the house, dishes, laundry, drop things off in the mail, pick up the kids, etc... When robots can accomplish these tasks they will become regular household items. I'm not big on having a maid in my house (can't afford one even if I wanted one) but I would love to not have to clean the dishes or do laundry. It robots could do that I would buy one over an iPad any day of the week.
      • by nman64 (912054) *

        I dropped thousands of dollars on the rings, fancy wedding, etc., and getting her to do any of those things was still much harder than building and programming a robot.

      • by lgw (121541)

        but I would love to not have to clean the dishes or do laundry

        Odd, I already have a dishwasher and a washing machine. Or did you mean a robot that would put dirty things into the existing robots that already clean them for you?

        I'd love a robot to iron and fold clothes, and one to clean my bathroom. And yeah, and iPad doesn't even come close.

        • Yes - I meant a robot that could do tasks that currently require a human. It's true that dishwashers and washing machines replaced human labor but they are simple tools / appliances. A robot is able to navigate from space to space while doing preset tasks - without requiring direct supervision. That's taking the dishes from the tables to the sink/dishwasher without tripping over toys or stepping on the cat.
          • by cynyr (703126)

            I wonder how such a robot would deal with my kids "pretend" dishes (really one of those cheep dorm room dishes sets)? I don't really need it to wash all of those.

            It would be cool to have a robot that would awaken in the middle of the day, pick up the house (this is much much harder than it sounds, unless every item has some sort of tag defining it's correct location), collect and do the dishes, etc and then go back to sleep. all the fuzzy categories and overlapping ones it makes it hard.

            • Yes. And I don't see it being too far-fetched for it to be commercially available in the near future.(10-20 years) Optics is there. Basic mobility is close. Processing power (well that's doubling every 18-24 months). But ... there's a lot more. I have no idea how close we really are to have a robot that can make the bed, pick up and take the clothes to the washing machine/dryer and then fold them and put them away.
    • Why not just use the same methods we currently use to train HR staff. In most places,HR is staffed by people operating as robots...

    • What "robots"'are you referring to? your dish washer or your roomba? Neither are "robots" in my point of view or what the fine article is alluding to. It's a about giving direction to a "multi-functional" piece of hardware in the true sense of the word "robot". Do this, do that, bend over etc. Not "clean under the tv" type of instruction.
      • Be careful what you say! My robot read your comment, and not it doesn't want to clean under the TV, pointing to your comment and saying "true robots don't clean under the TV."

  • I knew a Chinese girl who tried teaching a web-bot once. She gave up because instead of listening to her it kept trying to peak down her browse.

    • by mcgrew (92797) *

      You can't teach a robot anything. Robots are incapable of learning. Robots can't think. You program a robot. You no more "teach" a robot than you "teach" a computer. Even the IBM computer wasn't "taught" the facts it needed to win Jeapordy, those facts were programmed into a database. Watson does NOT think.

      And what does "peak down her browse" mean? That really looks like a racial slur by a semiliterate.

    • by Tablizer (95088)

      Damn, may scam has been exposed. Now I have to try to do it the old-fashioned way. I hate that!

  • Shut up and clean my floors!

  • "how do you teach humans to teach robots? Or, more precisely, how can you teach robots to teach humans to teach robots?"

    The real question is "how to build robots that can learn." Dogs can learn. Cats can learn. Birds and bees can learn. People ... the results are not so good ...

  • Before I can teach a robot, it must be able to feel basic emotions: pain, joy, disgust, etc. Until it is capable of associating environmental interactions into at least positive and negative stimuli, it cannot be taught anything... at least not in the same fashion we learn.

    All our basic learning functions are derived from emotional responses; And much of our learning is by observing others engaging in certain behaviors/tasks/activities and noting their emotional display (positive or negative). A robot tha
    • by geekoid (135745)

      no. It needs to have a definition for those things, doesn't need to feel them.

      Also, we have robots that can learn through observation. Don't really need meat emotions to learn.

      • by jd2112 (1535857)
        So we will have robots that will clean until hubbie leaves and then plops down on the sofa and watch soap operas all day.
    • by gmuslera (3436) *
      Emotions are the biological mechanism we use. Robots could have another kind of positive/negative reward system. Even recognizing the smile "gesture" from their human teachers could work as it. Anyway, trying to learning from speech, from natural languages, would be tricky, would be harder to associate semantics to natural languages than simulating a reward system analogous to emotions.
      • by mattack2 (1165421)

        Heck, robots recognizing smiles made it into popular culture over 10 years ago. I *think* it was this episode of Scientific American Frontiers, http://www.pbs.org/saf/1510/index.html [pbs.org] (Robot Pals), if not an earlier one. This actually may have shown a later version of the robot (the one that kind of looks like an Ewok).

        There was a creepy robotic face with exaggerated mouth that would recognize people smiling at it and making other faces, and make those faces back.

    • by blue trane (110704) on Friday March 09, 2012 @03:11PM (#39303873) Homepage Journal

      Do you really want self-driving cars to feel road rage?

      • by Oswald McWeany (2428506) on Friday March 09, 2012 @03:19PM (#39303983)

        Sometimes road rage is necessary. There is no other way to proceed at a busy four-way stop in South Carolina.

        See, in South Carolina- four way stops arn't what they are in the rest of the US.

        Rules here are. Stop. If there are any other cars nearby- the most aggressive driver proceeds next. Doesn't matter who stopped first- it's the most aggressive driver- (or the driver with biggest silly-wheels on their pick-up) that gets to go first.

        There are exceptions... sometimes a pick-up truck with just moderately big silly-wheels will trump a pick-up with 6ft wheels... but only if it is a police pick-up.

        If the driverless car had no road-rage he'd be stuck at the four-way stop for eternity.

        • Soon all cars will be driverless. Like, how many horse-and-carriages do you see at those South Carolina intersections these days?

          • Not many... they tend not to make tyres big enough for carriages- they get stuck at four-way stops for eternity.

            Yeah, I think all cars being driverless will be a big improvement. Maybe I will see it in my lifetime.

            It may be a while before ALLLLL cars are computer driven, but even before then- certain down-town areas of cities could be zoned "Driverless only" makes more sense in a city to have driverless cars so you can pack more in- have them navigate more efficiently. Have them talk to eac

        • by artor3 (1344997)

          Nah, eventually the car behind the driverless one will ram it forward through the intersection.

    • You appear to falsely believe that we are not machines. Surely, we are the bio-mechanical variety but our machinery learns (or fails to, if you'd like to take the pessimistic view) quite sufficiently, though with limitation.

      The knowledge representation and rule based approaches are inherently limited in the sense that you state. That is only a single portion of a much larger field. Would you state the same limitations in regard to the forms that enable your own perception and reaction, "emotional" or
  • It's a Roomba, it vacuums my floor on schedule and does a decent job, also drives back to the charging post by its self, other than cleaning its brushes it's entirely autonomous.

    • by WillAdams (45638)

      Excellent point.

      Count in the nifty automatic lawn mowers as well.

      Less mobile, but also arguably qualifying:

        - bread machine
        - coffee maker (if set on a timer)
        - 3D printer (printrbot)
        - CNC mill (shapeoko)

      I'm trying to decide which of the latter to get first (really wish there was a hybrid design where one could build one, then add the other functionality by simply swapping in some parts)

    • by Tablizer (95088)

      Yeah, but have you noticed you haven't seen the cat since Tuesday?

  • by geekoid (135745) <dadinportland AT yahoo DOT com> on Friday March 09, 2012 @03:02PM (#39303717) Homepage Journal

    Gets my mail, answers my phone it I want it to, looks for music I might like, l lets me know of up coming social event. Play chess against it.

    It's called a 'smart phone'. Hell, with minimal effort I can have it follow me on it's own accord.
    I mean, some Lego motors and gear, and it can us it's camera to follow me. Might take a day.

  • Robotics is dead (Score:5, Interesting)

    by Spiked_Three (626260) on Friday March 09, 2012 @03:04PM (#39303749)
    To some extent I disagree that dogs do not learn as humans. Its a reward and/or pain/discipline system and it works well on both. Robots on the other hand, may not feel reward or pain, so something new does need to be developed.

    I will agree with his disagreement on the first line. As a retired hobbyist I am hugely directly connected to the current state of robots. They are stagnant, dead, and useless with the exception of the Vacuum cleaner bots. There are some super high end stuff going on, but it is far more akin to smart remote control. Computers are not anywhere near fast enough at present. And as long as we stick to the von neumann model for their design probably will not be for 20-30 more years.

    But i have been an advocate of having moral discussion now, before it is too late. Saw a good short clip this am; http://boingboing.net/2012/03/08/disturbing-and-poignant-video.html [boingboing.net] about a robot that becomes self aware. Do we kill her, or let her be free? If we let her be free, what incentive ($$$) does anyone have to build her? If she is a slave (ie sold for profit) can we justify the treatment of any self aware being that way? And if so, why not retroactive?

    Anyhow, back top the subject; we need research in that area. Today for a current project I am looking at http://mnemstudio.org/path-finding-q-learning-tutorial.htm [mnemstudio.org] for ideas on what is going on now. I assume it the field will advance.
    • There are plenty of animals that are self-aware that we treat as slaves. Why would non-living robots be any different? "Young chimpanzees have outperformed human college students in tasks requiring remembering numbers." [citation] [wikipedia.org]
    • by geekoid (135745)

      How to you become a retired hobbiest? Isn't that just stopping to do the hobby?

      Anyways, robots are not stagnant, not by any stretch.
      Big dog, cheetah, we have bipedal robots that run.

      You need to define sentience. Please apply the definition to fire. If fire can be described with your definition, then the definition is wrong, OR fir is sentient.

      Oh, and we have system that can learn stuff we don't know and can't figure out when they tell us:

      http://www.radiolab.org/2010/apr/05/limits-of-science/ [radiolab.org]

      If you haven't r

      • You retire, and take up robots as a hobby. It isn't so hard really.
        I've had this discussion a 100 times. Most of the local robotics club consider black line following toys with 2 wheels robots. If that is your definition, then sure that type of robotics is alive and well. Colleges are also doing some cool stuff, and showing it with highly edited videos. Truth is none of them can go for more than 2-3 minutes and do continuous tasks without tethering. So sure, research is still alive. A robot to fetch me a b
    • Hobbyist robotics seemed to be stuck in the mid-1980s for several decades. The hardware got cheaper, but the systems remained about as dumb as 1980s industrial robots. Lego Mindstorms is an example. In the last few years, though, there's been more movement. Hobbyist robots are starting to use SLAM, vision processing, laser rangefinders, Kinect devices, and machine learning. All that stuff can be done on low-end hardware now. (At the $1000 level, anyway. We're not down to $100 yet.)

      There's now enough co

    • I will agree with his disagreement on the first line. As a retired hobbyist I am hugely directly connected to the current state of robots.

      As a current active researcher in robotics, I am also "hugely directly connected to the current state of robotics" and I have to say I disagree completely. I have seen amazing robots in person, doing things I couldn't even imagine robots doing 10 years ago. 10 years from now, I can imagine a world where robots are doing things I can't imagine today. Honestly, the only thing keeping robots out of the home today is cost. Robot vacuum cleaners are a notable exception that made their way into households because

      • by Sperbels (1008585)

        Honestly, the only thing keeping robots out of the home today is cost

        What useful tasks can they perform in a house right now besides vacuuming? I haven't seen a robot yet that can come into a house and do the dishes or my laundry without screwing it up completely.

        • by Carnildo (712617)

          haven't seen a robot yet that can come into a house and do the dishes...

          I've got a machine that can perform every part of the process except the "remove dirty dishes from the table" step.

          • by lgw (121541)

            Does it do the "put clean dishes in the cabinet" step too? Or did you take the batchelor approach of just declaring the dishwasher to be the place where clean dishes are stored, and leaving the dirty ones around until it's empty? Not that I'd know anything about that ...

            • Well, the solution to this is to have two dishwashers, one for the clean and one for the dirty dishes. As soon as the dirty dishes dishwasher is full, it's started and then takes the role of the clean dishes dishwasher- :-)

          • by Sperbels (1008585)
            Nor does it do the loading of the dishes part, or the cleaning of the dinner table part either....which is really what I'd like a robot to do...and if it looked like Lana Del Rey, that would be good too.
      • Mercedes is including more and more autonomous components into their production cars. It won't be long before it can completely drive itself.

        If that is the case, where might it decide to go and why would it want to go there?

      • The things you have seen are as I said, highly edited and trivial. It takes a car load of computers to follow a line, and in the end, is that a robot? You also Mischaracterize household robots; some (neato) use SLAM, sensors and laser scanners - they are closer than anything else out there. They don't just bump around (like iRobot does). The Kinect is a joke, although it may advance things accidentally, I agree. It brought commodity prices to an acceptable substitution, and thus many more people are able to
        • The things you have seen are as I said, highly edited and trivial.

          I'd love to hear your definition of trivial... and highly edited?

          It takes a car load of computers to follow a line, and in the end, is that a robot?

          I can make a robot that follows lines with a microcontroller. My lab built a car that can drive autonomously (which is incredibly more complex than simple line following if you didn't know) with a few mac minis. I can make a robot that can autonomously navigate outdoor urban environments with a single commercial laptop computer.

          The Kinect is a joke

          How exactly is the Kinect a joke? It's a 3D sensor with 10x the resolution as some of the best 3D laser sensors out t

    • I certainly hope so. Robotics has been a diversion from research into intelligence and consciousness. Humans have amassed an enormous quantity of knowledge--far more than any single human can absorb in a lifetime--but we still do not have a machine that can comprehend a bit of it.

      If and when we approach the creation of a system that can understand things, we will need to discuss the ethics of our relationship with such a system, but we have no pressing need to do so as long as we are merely building autom

      • by lgw (121541)

        It seems that people define AI as "the stuff computers can't do yet", then wonder why we don't have AI. Today we have every functional part of the requirements list for AI from the 60s. We have speech recognition, machine vision, expert systems, bots that can fool humans in turing tests, the whole checklist.

        In other words, we have all the useful bits of "AI", without having the part that would transform a robot maid from a tool into a slave. Why would you want that part?

        • In other words, we have all the useful bits of "AI", without having the part that would transform a robot maid from a tool into a slave. Why would you want that part?

          We have all the bits of "AI", except the intelligence. I contend that creating a mind which, potentially, is better informed than any individual and can reason more clearly and deeply than we can, would be of immeasurably greater value than all the tools.

          • by lgw (121541)

            We have all the bits of "AI", except the intelligence. I contend that creating a mind which, potentially, is better informed than any individual and can reason more clearly and deeply than we can, would be of immeasurably greater value than all the tools.

            What makes you think that an intelligence running on silicone would have those characteristics? As one AI researcher pointed out - there's no reason to even expect that an AI would be good at math. After all, intelligence is an emergent property of neurons, it's quite hard to go from "a neuron works like X" to "the brain works like Y" (in fact, we've never been able to do that from first principles AFAIK). Why would you expect to know anything about how machine inteligence works from knowing how a CPU wo

    • by rdnetto (955205)

      What are the ethical ramifications of programming them to get pleasure from serving us? If they don't want to be free (consider the house elves in the Harry Potter series), there aren't as many issues.

  • I knew a guy named Ivan and he's pretty out there. I am told that Ivan taught his nephew a little something.

    Every time he wanted a cookie he was to go up to his mother and say "Trust fund".

    I am wondering just how elastic these learning mechanisms will be, or will they really be pre-programmed routines where it basically learns to recognize your voice, and you can set a few parameters for which room is the living room or kitchen, but not much else.

  • by Anonymous Coward

    If my wife wanted a mindless automoton that could that could take instructions in natural terse language, and can be ignored the rest of the time, she would have gotten a husband. ..... oh, wait....

  • Sure, dogs don't learn the same as humans do, but that's because they have their own naturally-developed instinctive way of doing things. Robots don't learn the same as humans do because their technical capabilities are different, but the way they learn is still human design. The way a robot learns can be changed, and it will most certainly evolve to adapt to new technologies, but no matter what there will always be human design behind it all.
  • We need to backup a step and first look at how are we going to teach robot engineering students to teach robots to teach the human customers to teach the robots?

    Of course, Big Software will get involved and next thing you know, Robots will do things like, "I see your trying to mop the floop, can I be of assistance?" and then proceed to not only mop the floor but the inside of the fridge and the cat, while surreptitiously installing a company maid-bot in the upstairs bedroom.

  • by xclr8r (658786) on Friday March 09, 2012 @03:21PM (#39304003)
    1. Fetch a beer from fridge.
    2. Walk the dog to do its business.
    3. Wash dishes.
    4. Mow the Lawn.
    5. Clean House.
    6. Make robot programmable and able to share/sell programs ala app store.

    People don't want to program stufff and you are not going to change that behaviour. Just make their life easier and give the innovators the tools to accomplish this and the rest takes care of itself.
  • Three words: Global Thermonuclear War.
  • Depressed. (Score:2, Offtopic)

    by alvieboy (61292)

    Now I'm really feeling very very depressed.

    Marvin

  • I'm pretty sure I'm not a robot, but can someone teach me how to read that article... Look, when it comes down to it, it should have nothing to do with humans adjusting to how we can best teach a robot and adjusting our teaching methods to fit. It should be about designing machine intellegence to learn in a "human" way. If it's the former then no, "robots" as they say will not be predominant in the household. If its the latter and we design the "robot" to fit our needs and learn as we learn then yeah, they
  • I'd like to hear more people asking good questions...

  • Robot teach YOU lesson!

Science is to computer science as hydrodynamics is to plumbing.

Working...