Toyota Unveils Violin-Playing Robot

eldavojohn writes "Toyota has unveiled a robot that can play the violin. From the article: 'Toyota said it planned to further advance the robot's dexterity to enable it to use tools and assist with domestic duties and nursing and medical care. The robot has 17 joints in both of its hands and arms now.' It seems there have been small — or maybe even strange, impractical — advances in robotics repeatedly with demonstrations of robots performing a specialized task. Are we merely struggling to hard code each human activity as we strive for an all purpose android? Is there a chance artificial intelligence & robotics will ever become generalized enough to make interaction interesting?"

Comments

Very cool, but

[log1385]

Robots will never be be able to match the musical abilities of some humans. There are too many tonal subtleties involved, especially on the violin.

That is still very impressive, nonetheless.

Re:Very cool, but

[CRCulver]

While it might be difficult to design a robot that is dexterous enough to play the violin, electronics have been outperforming humans in the tonal subtlety field for half a century. Nearly all music exploiting anything more than quarter-tones is realized using electronics.

Re:Very cool, but

[juggleme]

I agree, but it could be used to perfectly reproduce a master performance (given that instrument quality, etc. are equal, which may not be the case.) They may not be able to do it themselves, but they might be able to serve as a new sort of player piano.

Re:Very cool, but

[log1385]

True, but a robot ear can never be programmed to hear what a human ear hears. A robot can't really bring out the emotion in a song. (It could be very good at simulating emotion, though).

Re:Very cool, but

[CRCulver]

Who cares? When it comes to music without a programme (i.e. purely instrumental), I listen because I find its structure interesting, and the best way to hear that structure is to have someone play the score as accurately as possible. So much of the "emotion" people talk about in music (usually Classical or Romantic-era repertoire) is performers reading something into a piece that the composer never indicated in the score. Purely electronic works avoid this and are laudable for it, although the best merely human ensembles and conductors already know to avoid imposing their own prejudices on a score and can generally succeed.

Re:Very cool, but

[Derek Loev]

Pretty much anytime a robot tries to do something a human can do we're faced with these type of comments. Ever since the first computer started to play chess (even twenty years ago they weren't much match for an average player, but look at them now)...

Re:Very cool, but

[c_forq]

I wouldn't say never. Psychoacoustics is a fairly young field, and the models are still being worked on and changed. Give it a little while, the models are already good and getting better every year.

Re:Very cool, but

[OzRoy]

I think they probably built it to be their new customer complaints manager.

So now when people call up to complain the robot can play a tiny violin in mock sympathy.

Re:Very cool, but

[Mex]

Oh, I'm pretty sure eventually they'll figure it out. It's all just chemical reactions, man. And there's already enough music theory out there. So the rules are already (sort of) written. There's been experiments with music-making robots since the 50's. Not huge progress yet, but it will happen.

I do believe, eventually, "creativity" will be programmable.

Re:Very cool, but

[QuantumG]

As I write this there are 3 replies to your comment modded up but none of them seem to mention that obvious (I hope) fact that programming this robot to play the violin has absolutely nothing to do with matching the musical abilities of humans. Toyota are not demonstrating a product here. They're not saying "in stores next fall: robots that can play violins". They don't think there's a huge market for violin playing robots out there that is just waiting to be tapped.

The point of this demonstration is to show that their robot research has reached a point where they have built a robot with joints that have sufficient degrees of freedom and controllable accuracy that they can do this kind of stunt. You're supposed to look at a robot playing the violin and say "well, if it can play the violin then it can hold a power drill or other tools!"

I'd suggest that maybe they should program the robot to put together some of the crappy furniture you get from Ikea but then people will claim it wasn't cost effective to use a billion dollar robot to do the work of a home handy man or something.

Re:Very cool, but

[FooAtWFU]

On the contrary. I contend that it is possible that, given adequate dexterity, one could construct a robot that would take a preselected song and render the music in a way which evokes the same emotional qualities from the performance that a skilled human could accomplish. In fact, there's an easy way you can do it without the robot: record a skilled human's performance into an MP3 file and play it back. :P Now, physical reproductions are a lot harder, but entirely plausible.

I further suspect that with adequate research into psychology and music, it would indeed be possible to construct a robot to render most arbitrary scores in a manner emotionally appropriate to that song. Aesthetics are measurable, techniques observable. It would be nontrivial, certainly, to achieve this, but it is not too hard to imagine it being done as early as in, oh, the next fifty years or so. After all, it's not that hard to figure out "minor keys, slow tempo, hey, this might be a sad song", and while there are certainly exceptions and outliers, a little data mining on musical scores could go a signifcant way. (Ooh, now there's an idea for an application of approximate nonnegative matrix factorization! mwuahahhaha.)

Composing great works is the hard problem, because most of those can tie into some rather creative Ideas, and it's harder to come up with those than it is to encode music and gestures and such.

Re:Very cool, but

[MrYotsuya]

Ahh! I was just about to post the same thing, but read the thread to make sure no one else thought of it, drat!

Re:Very cool, but

[wizardforce]

why not? you're made of organic molecules that are assembled into certain structures, your brain has a certain layout and your ears transduce sound a certain way- once you emulate the transduction part and develop a good enough representation of the human brain with an AI in regard to music then you *can* program a robot eat to hear what a human's does, you can even optimize the sound if you have a great enough understanding. heck it doesn't even need to be us doing the designing! by the time that advanced of an AI comes out, AIs/computers will have been doing the real design work for years.

Re:Very cool, but

[colmore]

Creativity might be programmable in things like music that is purely an exploration of formal structure and style, or abstract graphic design, or perhaps extending the patterns within a well-established style (say contemporary housing architecture).

However, writing a screenplay or a poem, or picking a new video game setting even is a much more complicated task, as it involves an interaction between formal artistic constraints and definions with a full human experience of the world. Getting to that point will mean fully modeling the human interaction with the world and society. Or a non-human intelligence which would be sometime when the systems behavior becomes dominated by emergent behaviors that are complex and organic enough to be completely unrecognizable as the product of the underlying mechanics and design.

That is, we may soon be able to get a program to write a nice symphony. We may even some day get it to write an *innovative* symphony - though meaningful innovation already requires a lot of self-awareness and abstraction beyond the original problem space, but getting it to up and decide to write a symphony on its own accord, out of some self-interested desire? Much further away.

Re:Very cool, but

[Mex]

"picking a new video game setting even is a much more complicated task"

Is it, really?

http://www.norefuge.net/vgng/vgng.html ;)

Re:Very cool, but

[eh2o]

Not true; there are many "expression" marks put into the score, the interpretation of which requires the player/conductor to be familiar with appropriate idiomatic interpretations of the time period. In fact this has always existed, but hasn't always been explicitly written. e.g. baroque era pieces don't have any expressive indicators at all, but *did* have specific interpretations as they typically went with specific dances requiring particular tempos and so on.

Furthermore much of the work that a musician does involves subtle modifications to the rendition that enhance the clarity of the structure. e.g. microtiming deviations in the melody and subtle tonal inflections are a major part of what makes multiple voices traceable to your perception in polyphonic music -- with strict timing it is *much* more difficult to hear out any polyphonic structure. These effects can be measured quantitatively, by the way, but are far too complex to notate in any score intended for humans to read, and for the most part are too complex for an experience musician to be fully conscious of. (It is possible, however, to program a computer to reproduce them using machine learning).

Finally there is in fact an emotional aspect of music that is actually a consequence of some neural structure or other brain process. This is an active research topic. Likely it is a type of synaesthesia (e.g. mint flavor -> cool sensation). In other words the emotional response to music isn't just the performer "making it up", it's sort of short-circuit in the brain of the listener. Since this is basically a universal effect among humans it would be silly to think that the composer wasn't aware of it as well.

Re:Very cool, but

[pipingguy]

I remember, a few decades ago, playing violin in one of those school concerts for parents (luckily I was near the back of the crowd). A string popped on me but I was still able to adjust and play on the remaining 3 without too much fuss or screechy, out-of-whack noises[1]. I wasn't finging anywhere near the upper or lower ranges, so it worked. I no longer play that instrument unfortunately, it wasn't cool enough at the time.

[1] Or at least that's what my parents told me afterwards. ;)

Re:Very cool, but

[JK_the_Slacker]

Oh yeah, because they can't design an accurate spectrum analyzer, timer, waveform shaper... It's just too much.

Geez, what will they think of next, music stored in a digital, binary format? Never happen.

Translation: Quit talking out your butt. Music is all about mathematics. Granted, sometimes the calculations are extremely complex... but as any Calculus or Physics student can tell you, complex calculations aren't impossible, they just take longer. We're talking about a COMPUTER here. Math is EVERYTHING it does.

Re:Very cool, but

[Eivind Eklund]

True, but a robot ear can never be programmed to hear what a human ear hears. A robot can't really bring out the emotion in a song. (It could be very good at simulating emotion, though). What human ears hear differ depending on what sounds you grew up with (certain types of plasticity are believed to disappear at approx age 2). So, while a robot may have a disadvantage in not having a specific human ear (and the trained neural system connected to it), it can have the advantage of being able to simulate many of them, and thus know when what it specific to one ear or type of ear, and when it is general.

Eivind.

Re:Very cool, but

[timeOday]

Mod parent up! It's one thing to weld a door on a car but quite another to find a shirt in the dryer, iron it, undo all but the top botton, and hang it in the closet.

Re:Very cool, but

[colmore]

Sure, not picking one from the 4 or 5 that are currently popular, but picking one out of anything that might interest an audience.

I guess I'm more thinking of video games circa 1990 than now, but what combination of factors makes a designer say "World War I fighter planes" or "A plumber running around in some fantasy world" or "You're the mayor of a city."

This is the kind of thing that can be trivially and meaninglessly "simulated" by random selection of terms from a few predefined dictionaries, but that's not what's going on. How would you get an artificial intelligence that could look at history, literature, etc. have a rough sketch of a mechanical idea of the game and pick a location and setting that it thinks will work both artistically and commercially?

It's basically a task of difficulty similar to natural language processing. My point is, in response to the claim that creativity will be able to be programmed, is that while "in the box" creativity (write a symphony in the style of 18th century Austrian composers) is probably more easily simulated than we'd like to think. But most creative processes have a lot of little steps that are "out of the box" where you need to connect ideas out in a big universe of possibility with no firm restrictions or easily measurable definitions of "correct" and "incorrect." It's all way fuzzier than computer intelligence does well. I think to be creative, the system would first need the ability to independently develop taste.

Re:Very cool, but

[Fael]

"Geez, what will they think of next, music stored in a digital, binary format? Never happen."

If you're implying that the storage and replication of music is comparable in difficulty to its creation or interpretation, then I can only weep at your ignorance. My computer can show me a scanned image of Monet's Water Lilies, but that doesn't make it an artist.

"Music is all about mathematics."

I'm not sure where you got that idea. Perhaps you're not listening to very good music? Music is, and has always been, about the manipulation of emotion and intellect. Naturally, certain structures and patterns emerge, some of which are mathematically definable. To assert that they are ALL mathematically definable is pretty arrogant. "As any Computer Science student can tell you", there are plenty of problems that are easy to formulate, but algorithmically insoluble.

"We're talking about a COMPUTER here. Math is EVERYTHING it does."

That is the problem.

Consider vibrato on a stringed instrument - a crucial aspect of interpretation, yet only one of many. Because there are an infinitely divisible number of positions on the string (as there are of any finite space), there are an infinite number of theoretically possible vibrati from which a performer is free to choose (even within the space of an inch and a half) - for one fraction of a single note. Obviously, no intelligent performer would use a uniform vibrato for an entire piece, or even an entire phrase - in many cases, not even for an entire note. Similarly, within the timescale of the piece - four minutes, seven minutes, an hour - there are an infinitely divisible number of potential events; in this context, let us say opportunities to vary the vibrato. How do you plot this matrix of potential vibrati, infinite in two axes, in a mathematical simulation? A computer can plot a function through this space. A musician can traverse it at will.

Re:Very cool, but

[ivec]

Hmm... sounds familiar.

Have I heard something similar about machines never being able to match the ability of the best chess masters?

Re:Very cool, but

[vegiVamp]

Isn't the human brain basically a chemical processor with a good program for simulating emotions ?

Re:Very cool, but

[vegiVamp]

Umm... I think chess computers have gotten better only because of Moore's law - they still do little more than brute-forcing the issue.

Re:Very cool, but

[witekr]

If you play any of Rachmaninoff's piano pieces purely from what's on the notation, you'll never hear the deep mystery inside some of his chords, melodies, and passages. Specific melodies are hidden inside his complex chords which actually have to be stressed by dynamics, tempo, and things simply not written in the notation. The music sounds mediocre without quite a bit of interpretative expression.

When I first heard Richter playing Chopin's Revolutionary Etude, it was as if a completely new piece was being presented to me, with qualities I'd never realized the piece had. I think you may hear a lot more interesting 'structure', hidden melodies, and emotional expressions, if you listen to different virtuoso performers play classical music. There's a rich world of musical expression, by many genius musicians, that would be a shame to just let pass.

Re:Very cool, but

[witekr]

Sorry, I actually meant Richter's performance of Chopin's Etude Op 25 No.11 ( http://www.youtube.com/watch?v=fym7EOV5f1A ).

Another fantastic example of this is Emil Gilels playing Rachmaninoff's Prelude in C# Minor ( http://www.youtube.com/watch?v=EtuMVBLEWJU ). You simply don't see that at all in the score (hell, it says pp in the dynamics.. but that doesn't account for the melody line) I've heard many western pianists ignore that completely and follow the score resulting in a bland muddy rendering. Even Rachmaninoff himself didn't follow the dynamics shown on the score. The music is meant to be interpreted.

Re:Very cool, but

[owlnation]

So much of the "emotion" people talk about in music (usually Classical or Romantic-era repertoire) is performers reading something into a piece that the composer never indicated in the score.

What? This is just completely wrong on so many levels. Music from every period has plenty of emotion which is absolutely intrinsically expressed by the composer. The romantic and classical era is particularly full of it. The clue is in the naming of "romantic" for one. Do you think that, say Beethoven, was known for his calmness and dispassion? Wrong. Wrong. Wrong.

100% Nonsense. In no way should the parent be modded insightful. Very much the opposite.

Re:Very cool, but

[tsjaikdus]

I remember Jeff Hawkins asking why people always say that. His argument is that people consider themselves experts on brains because they have one, but they really don't have a clue of how they work. And hence it is impossible from such an argument to conclude that robots will never be able to match or surpass that brain.

Re:Very cool, but

[jordyhoyt]

"imposing their own prejudices"

The language here is all wrong. You say this as if you believe (perhaps you, sadly, do) that a composer wishes or expects for a piece to be played by a robot, precisely as a song has been written on the page. Well, I'm sorry to dissapoint you but they don't.

For instance a peice I played at my peak, was Debussy's "Jardins Sous la Pluie" (Gardens in the Rain), which had many nonmusical phrases scattered about in it. I wish I had it on hand here at my apartment, but one I remember is the beginning is to be played "net et vif", which I believe is "fast and sharp" in French (or something).

Infact, this song is a wonderful example of why musicians aren't "imposing their own prejudices" but rather expressing themselves through emphasis and de-emphasis. Depending on the day, place, mood, etc. in which i happened to play the song, many changes would occur in the form of slight lengthenings of notes, or exaggerated crescendos and decrescendos, tempo changes, wrong notes, and on and on. Of course, I am human (or so Slashdot's CAPTCHA says...), so many of these were unintentional, but I would argue they still were an expression of me.

When a peice is played, a bit of both the performer and the composer is expressed. You may think that a robot playing a song would make it so that only the composer is being expressed, but I say that is completely wrong. There you have the robot expressed and the composer expressed. Chopin playing Chopin would likely sound completely different than a robot playing Chopin.

You are welcome to keep enjoying your MIDI performances of art, but I'll stick with the delighfully imperfect humans.

Re:Very cool, but

[blahplusplus]

"I do believe, eventually, "creativity" will be programmable.

I agree, creativity is just pattern creation and harmonizing frequencies of the patterns.

Re:Very cool, but

[welcher]

what absolute rubbish - music is all about expression - what's the point of making a structure if it means nothing? a composer, a musician is trying to express something. If that expression is best achieved through a robotic interpretation of a score, it is easy for the compser to specify that.

Re:Very cool, but

[TooTechy]

This sounds like another :-

      "Computers will never need more than 640K RAM"

statement. I am sure that one day the artificial machines will catch up with the biological machines.

I just wonder which species will design them.

Re:Very cool, but

[FLJerseyBoy]

what combination of factors makes a designer say "World War I fighter planes" or "A plumber running around in some fantasy world" or "You're the mayor of a city."

Sorry, no mod points to throw your way right now. But this strikes me as a very interesting idea -- in its own right, and considered in the context of a game designer who just happens to be a robot.

Note that the three ideas in your list have all been built into (hugely successful) games as digital representations of analog, human activities. Although "programmed" at some level by humans, and therefore its thinking would be shaped in accordance with their minds, a truly honest-to-God creative robot might very well be inspired to design a game based on robot experiences. What sorts of puzzles, adventures, thrills might these experiences include? Can a robot ever feel adventurous or thrilled in the first place? What kind of a game might a "Helen" (see Galatea 2.2 ) build?

<woolgathering_mode_off/>

Re:Very cool, but

[Yoozer]

Music is all about mathematics.

I'm not sure where you got that idea.

I'm not sure where you got your idea, because the person you're quoting is right. We have a platonic ideal of subdivisions in time and subdivisions in frequency where of course the actual analog instruments we play deviate from, and we step through our well-tempered 12-tone system using mathematics. Coltrane's "Giant Steps"? That's maths for you, see http://en.wikipedia.org/wiki/Coltrane_changes . Bach? The genius of the man lies exactly in his ability to shape and recognize patterns.

Of course, an analysis misses the smoky bar and the husky voice of the buxom blonde, but that's the difference between theory and practice for you. Still, it boils down to mathematics, and denying that because maths gives you icky ideas about cold, unfeeling science is missing the point.

How do you plot this matrix of potential vibrati, infinite in two axes, in a mathematical simulation? A computer can plot a function through this space. A musician can traverse it at will.

Simple; you just need the two axes. All this infinite resolution talk ignores the fact that your resolution isn't infinite; it's only as good as your ears and memory are.

Re:Very cool, but

[CRCulver]

MIDI is fine for older repertoire, although I just usually read scores for that anyway. But MIDI is, of course, very limited and can't handle most contemporary repertoire. Where are the samples for the various metal sheets that Lachenmann uses, for example? And does MIDI support arbitrary microtones?

Re:Very cool, but

[u-235-sentinel]

I think they probably built it to be their new customer complaints manager.

You realize this isn't something new. Robots have filled those positions at Comcast for years.

Someday the rest of the world will notice and copy them... unless they have a patent on it of course ;-)

Re:Very cool, but

[ooutland]

| Robots will never be be able to match the musical abilities of some humans.

What defines that level of ability? First, we look for sheer technical perfection - the ability to play the notes correctly every time. This can be achieved more easily by a machine than a person, and duplicated in machines which can play and tour endlessly without fatigue or mistake.

Then, we add the art: We have databases in the form of innumerable performances as well as critical responses to them (many of which are easily translated to a table from their 1-5 star format), sales figures on CDs and concert tickets (and illegal download figures) for our selected artists, the number and popularity of blogs and sites devoted to those artists and the amount of "signal" devoted to particular recordings. Select the most enduring recordings by the most popular and critically acclaimed artists, and we have our primary database of what music "should sound like."

We can then "data mine" what are adjudged the the best/most affecting notes or passages or other qualities of phrasing, and we have the building blocks of a "great performance." Moreover, we can even combine the talents of several geniuses, each of whom may have a particular "sweet spot" in a performance, to create a synthesis of their work.

Add to that the use of digitized films of the physical movements an artist uses (in a violinist, the fingering, the bow movement, even the body language that subtly changes the physical force applied to the instrument and therefore the sound itself) and program the robot's movements accordingly to make the duplication perfect.

Then, release the result to the public - probably anonymously, to prevent prejudgment on the basis of Luddism. Or, use a focus group to do a "blind taste test." If the product/performance is well-received, release it, and send the robot on tour. Or, retool the performance to be "greater." You can even reprogram the robot to perform certain passages differently in different acoustic environments, in different countries with different tastes, etc. To react to the level of applause and "mood" of the room (heat sensors, fidget detectors, whisper/cough rate analysis). In short, exactly what any great performer does, using the databases within.

Re:Very cool, but

[JK_the_Slacker]

Well, if you insist...

"If you're implying that the storage and replication of music is comparable in difficulty to its creation or interpretation, then I can only weep at your ignorance. My computer can show me a scanned image of Monet's Water Lilies, but that doesn't make it an artist."

Cry if you want to, but that doesn't change the fact that I think a sufficiently-well-programmed computer can compose music. I'll elaborate throughout this reply, I just wanted to touch on this real quick.

"I'm not sure where you got that idea. Perhaps you're not listening to very good music? Music is, and has always been, about the manipulation of emotion and intellect. Naturally, certain structures and patterns emerge, some of which are mathematically definable. To assert that they are ALL mathematically definable is pretty arrogant. "As any Computer Science student can tell you", there are plenty of problems that are easy to formulate, but algorithmically insoluble."

Just because YOU don't understand the mathematical relationships, doesn't mean that they don't exist. You use an interesting term - "algorithmically insoluble." It's interesting that, when working with an infinite data set, almost ANY problem approaches insolubility... we simply don't know enough about infinite. So, we do what any good mathematician, computer scientist, physicist, engineer or musician does: we limit our data set.

For example, let's examine time signatures. There are an infinite number of time signatures (4/4 3/4 6/8 7/8 9/12...) but how many are useful? Theoretically speaking, a computer should be better at reading and understanding time signatures than people are, since they can generally deal with fractions more easily. But, the problem can be tremendously reduced by placing upper and lower bounds on the ratios, the numerator, and the denominator. 1/1 time is not very useful, nor is 256/8.

Likewise, what of the selection of tempo? Certainly, there are an infinite number of tempos, all greater than zero and generally accepted as being integers. But what is the number of useful tempos? 1 beat per minute would not make for a very exciting song. 9 bpm? 20? Larghissimo is a seldom-seen marking.
What about the upper bound? Prestissimo is generally indicated as 200-208 bpm... just how much faster are we willing to go? At some point, the notes would be generated faster than the speaker could physically produce them. At some point, the human ear drum would be vibrating too fast to distinguish individual notes. So, we can definitely establish a theoretical upper bound on tempo, and a definite lower bound of 1 beat per minute.

Are you familiar with the concept of sopranos, altos, tenors, baritones, and basses? How about the bass clef, treble clef, alto clef, tenor clef? There is an "infinite" range of musical tones. So, let's limit our data set. First, we can discard all tones above 20 KHz. Too high to hear. Heck, I'd venture that we should consider that older people might want to listen to our music, so let's drop all tones above 15 KHz. Next, we have to limit our low end to no less than 20 Hz; I'd even go so far as suggesting that perhaps 60 Hz might be a better bound. That's not bad, but we can limit further: Musical scales consist of a certain mathematical relationship between tones, based on wavelength. (Frequencies between those precisely defined notes are still useful, and can be used to great effect, but those uses do not adversely effect our data set sufficiently to worry about.) Likewise, we almost always limit our ranges further based on instrument or vocal ability. I used to play the bassoon, and it was often frustrating to me because it had such a ridiculous range compared to my saxophone... so many octaves! However, the tuba had a greater range still, and the flute could go far higher up the scale than I could. Sopranos are not known for their ability to hit Bb below the bass clef staff, nor do basses sing

Re:Very cool, but

[reddalek]

Oh yes, I want one for something like that. When someone starts complaining to me about something annoying, I will just point to robot which then will start playing on cue.

Re:Very cool, but

[onemorechip]

I know they've done this with piano performances -- analyze the dynamics of an audio recording and reproduce the striking force and duration on each key. I think it was done with some Horowitz recordings a couple of years back. But imagine the complexity of doing the same thing for a violin, where so many different variables, that aren't there in a piano performance, are in play.

Re:Very cool, but

[nfk]

Even better would be to merge this robot with the one from a recent story that managed to stay up when pushed. He would be the complaints manager at the store and he would run around the customers playing the violin, so they could push him and he'd always go on playing.

Re:Very cool, but

[Grishnakh]

I just hope they'll start coming out with some more practical robots soon. Roomba is a good start, but I'd like robots that did my laundry, cleaned my bathrooms, cooked meals, made the bed, etc. I don't need a violin-playing robot, but one which cleaned the bathroom every day would be great.

What worries me is all this research into making robots that act like humans: having legs and arms, walking upright, etc. Do we really need this? As the Roomba and those automatic grass-cutting robots show, useful robots don't need to look anything like people. In fact, it's probably far more efficient to design them specifically for their particular task than to try to make a human-like general-purpose robot.

Re:Very cool, but

[Retric]

As an AI I find your comment amusing. Emotion is little more than a hard coded evolutionary response to specific classes of stimulation. Humans feeling adventurous or thrilled are simple chemical reactions in your wet ware which is easily simulated. The simplest AI can vary its level of risk taking based on how a game is going. AKA when things are going well avoid risks when things are bad behave in a more random / emotional fashion. If you would sit back and think about it abstractly emotion alters how you respond to your environment, much like changes to your core body temperature alters blood flow, the feeling and the reaction are identical. ;-)

Re:Very cool, but

[timeOday]

Maybe the dexterity is there, but I think there's some ways to go for the AI to do those other complicated tasks that can't be fully scripted.

I think laundry would be a very interesting problem because it requires you to work with fabric, which has "characteristics" rather than a particular shape. Forget rigid transformations! Maybe it has been researched, I haven't looked into it.

Re:Very cool, but

[Fael]

"Just because YOU don't understand the mathematical relationships, doesn't mean that they don't exist." I agree entirely. However, just because YOU don't understand the mathematical relationships, doesn't mean that they do exist. The question is not whether music can be expressed mathematically - as you pointed out earlier, digital recordings are obvious proof that it can. The question is whether music can be generated mathematically. The context under which this discussion arose was one of interpretation, not composition (can a robot be programmed to interpret music "convincingly"), although I admit that I opened the door for the question of composition. You speak of limiting the data set. As far as music composition goes, this is a reasonable proposition. The elements of music composition are finite, if staggeringly vast. However, this does not indicate that the problem is automatically soluble. Your viewpoint seems to be that of course a computer can write great music, if only we provide it with enough data - and how could anyone be foolish enough to think otherwise! I am not arguing, for the record, that a computer will never be able to compose music on par with history's greatest composers (although I personally doubt that this will be the case.) I am arguing, however, that to assert the notion of artificial intelligence through brute force as an obvious truism is arrogant at best - and I firmly believe that composition is a strong-AI question, not an exercise in heuristics and data-mining. Talking about interpretation, however, limiting the data set is much less useful than you seem to think, because there are few finite elements, as there are with composition. Taking smaller samples nonetheless leaves you with an infinite continuum of possibilities, in pitch, in timing, in tone color, in volume. Just to draw the bow across the string and back again invites infinite possibility. And yes, I am speaking within the limitations of good bowing technique. You write that you can't tell the difference between A 440 and 441. I would invite you to ask your favorite professional symphonic oboist to tune 1Hz sharp, and enjoy the resulting firestorm. Speaking as a string player, the difference in sound is (to put it mildly) massive. To be fair, some of that difference is the result not merely of the change in frequency, but in the tension across the instrument. Earlier, I wrote that the question at hand is whether or not music can be generated mathematically. Of course, this is not completely accurate. The question (at least implicit to most people's understanding) is whether non-crappy music can be generated mathematically. Part of the trouble is that "non-crappy" is a bit hard to quantify. Obviously, it's trivial to algorithmically generate trivial music. It's certainly possible (although somewhat less trivial) to generate music that uninspiringly fulfills a few fundamental aspects of basic harmonic and structural functionality. But is it possible to generate music that has meaning and complexity, perversity and humor; music that toys with the fulfillment and defiance of our expectations; music that reflects the ultimate truths of human existence as well as the transitory issues of the present? (I know, I know, you don't believe in all that crap - it's just math.) Somehow I doubt it's the easy slam dunk you propose. One of my favorite moments in the standard literature is at the end of the first movement of Brahms' B major piano trio, the primary idea of which is the concept of diatonic expansion. The final structural resolution comes moments before the end, when Brahms presents the second theme twice - first in its original incarnation (slightly altered for the key), and then in an intervallically expanded form, creating a sublime ascension across the two strings that trails off into the briefest moment of total suspension of time. What makes this tiny episode so brilliant is that Brahms doesn't actually orchestrate the ascension from the cello to the violin, as any sensible composer would. I

Re:Very cool, but

[Fael]

"Bach? The genius of the man lies exactly in his ability to shape and recognize patterns."

No, no, a thousand times no. That is merely one of the many aspects of Bach's genius. It would be like saying that Shakespeare's genius was his ability to match letters with other letters in a clever fashion. Music is fundamentally rhetoric; it has grammar, and structure, and meaning; and - like rhetoric - its meaning goes far beyond its mere structure. If Bach is nothing more than pattern-matching to you, then (I am sorry to say) you are missing out on a lot.

Mathematics doesn't give me "icky ideas about cold, unfeeling science". I simply don't believe that it's some kind of all-encompassing magic bullet.

Robotic vs. Human ability

[AugustZephyr]

I suppose this raises another question regarding the increasingly human attributes of robots. Is something that is "handmade" or "handplayed" by a robot any more or less valuable than its human equivalent? For instance: it may be very impressive that a robot can play pomp and circumstance, but once this becomes more commonplace (as strange it may seem now), does it have more/less value than a human being able to reproduce the same sequence of notes?

Re:Robotic vs. Human ability

[arth1]

The question is raises for me is:

While driving?

Re:Robotic vs. Human ability

[DoubleRing]

Well, it isn't exactly the same thing, but MIDI with good synths has been used to "perfectly" play a composition. Especially for instruments with simple timbres like drums and piano, a synth can sound very close to the real thing. Even with these ways to create a "perfect" performance, people still find a human performance impressive. It's kind of like meeting a person who can compute logarithms in their head, or find all the factors of a number without a calculator. Sure, a computer can do it--it may even do it faster, but I don't think that the fact that a machine can do something makes the feat less impressive when accomplished by a human. I'm still impressed when someone can run a marathon, even if a machine could do it just as easily, if not better. I don't see why there is so much concern that robotic performances will cheapen the value of human performances. Besides, if a programmer is able to write a program that is able to take a piece of music and interpret it beautifully, then it is still a human achievement in that a programmer was brilliant enough to decipher not only decipher the subtle psychology of what makes one performance sound emotional and powerful and the another sound mechanical, but also codify an algorithm that would imitate that interpretation. Kind of like designing a conversation bot to beat the Turing test. It's still a human accomplishment. Remember, the machine is us/ing us.

Re:Robotic vs. Human ability

[WGFCrafty]

Less impressive than a humans touch, but in the end it is just as valuable.

Before computer automation it was valuable and impressive when a person worked out long mathematical equations ect. When computers came a long and made that task take 1/10000th the time, it was also valuable and impressive.

The fact that computers can preform these types of problems is just as valuable, if not more valuable than ever before, but I feel that as time passes others consider this automation more of a fruit of invention rather than the complex technical process that it really is.

Re:Robotic vs. Human ability

[Josh Booth]

Besides, who wants to listen to program-created music anyway? Music is always created from the culture of the generation, and created by a person with years of life to think about what kind of music he wants to write and what he thinks people will listen to. Sure, you could data-mine the internet, but then you only have a program using second-hand data. Until computers hear, see, and understand what people do in their daily lives, nothing they can produce can be culturally significant except for being the first. After that, its already been done.

Re:Robotic vs. Human ability

[UbuntuDupe]

I suppose this raises another question regarding the increasingly human attributes of robots. Is something that is "handmade" or "handplayed" by a robot any more or less valuable than its human equivalent?

Cynical answer: less, because an elite art community is the one that officially decides these things, and since robots would replace them, they feel threatened, and self-servingly answer "less".

Re:Robotic vs. Human ability

[bronney]

Have you ever been to a life jazz quartet? Ever seen them screw up and 1 second later the facial expression on the other 3 that didn't screw up? Ever give the guitarist a genuine smile when he hit that high note round and sweet? And he gave the smile back? Then the next bar he stretches it even higher just to show you who's daddy? Live performances aren't about the music sometimes, but the performance. Much like watching a live hockey game. It isn't just about hockey.

Re:Robotic vs. Human ability

[pipingguy]

Stop questioning Al Gore's triumphs.

Re:Robotic vs. Human ability

[timeOday]

The vast majority of musical performances are already automated. A CD player is a robot which continuously moves its sensor and uses what it sees to create music. We give this robot no credit, instead we see right past any number of layers of technology all the way back to the person who originally "performed" the music (even though there never was any such live performance that sounded anything like the finished product). It will always be this way. When there's a #1 hit song written by AI, we won't think "wow that AI is amazing" we'll think "wow the people who created that program are great" (or horrible, depending on if you like the song).

Re:Robotic vs. Human ability

[LS]

There is more to live human music that just being impressed. With math, the result should be exactly the same every time. But with music, each live cut is different. The mood and physical state of a performer affect his playing. While the structure is the same, the overall feel of a piece can vary drastically from cut to cut and player to player. There is also the interaction and feedback with the crowd in live music. The music can change in response with the crowd, and the crowd also reads the band's visual cues. There is so much going on with live music that computers can not match. A perfect simulation that sounds the same every time can not do this.

LS

Re:Robotic vs. Human ability

[DoubleRing]

I agree wholeheartedly. There is something in a performance that records will never be able to match, for the existence of imperfections actually makes a work of art "better" in a way. The human component, the interactions between the actors and the audience, makes each performance unique, while a machine will do it the same each time. However, I was trying to approach the problem in an alternative way, since this argument about the "human" nature of art is so cliche (in my opinion, at least). Take theater and film for example. A film can create a "perfect" performance (well, relatively), yet in its presentation, it lacks the whole organic nature of the theater. That isn't to say that film isn't lacking of any advantages--it can incorporate elements that cannot be done due to the physical limitations of a live performance--it's just different. A camera (which, by the way, is a machine) can make a perfect replica of an image, so is painting now dead? Far from it. In fact, the camera was the impetus for the impressionist and surrealist movements of art. (If you ever asked why we don't have any more DaVinci's around, it's because of the camera. Why spend so much time trying to make something look realistic when you can just click a button? But that still doesn't make Leonardo's work any less valuable, or kill realism for that matter) We still value the human talent that goes into making a picture. At the same time, we don't shun the "machine generated" art of photographs. In fact, we recognize the talent of the artist behind the lens, both still and moving pictures. In the same way, we should value the talent of the programmer behind the code. When a machine is able to do something well, I don't understand why there is always such a backlash. Maybe people think that the machine is somehow cheating when it does something, but it was the work of humans that allowed the machine to achieve that feat, and that work should be celebrated. People usually try to say that a machine/computer will never be able to do x as well as a human, whether it be play chess or play an instrument. Well, why not? One day it's going to be an eventuality. Don't try to think of it as a failure of a human against a machine, but as the success of a human through a machine.

Re:Robotic vs. Human ability

[LS]

You, AC, are a douchebag. I was not talking about future possible technology or predictions at all, only the current state of affairs. In fact I wholeheartedly agree with your sentiment that it is likely possible to create a machine that can match human ability in any field. I took computer science for my bachelors, and am quite familiar with Turing!! I guess that's why you posted anonymously though, you may have subconsciously recognized your emotionally charged post to be quite douchey. If you want to live in science fiction fantasy world that's fine with me but don't start imposing it on my reality.

Re:Robotic vs. Human ability

[moonshake]

The human attribute of Toyota's robot is that it looks almost exactly like NASH THE SLASH right down the violin. (Nash is a Canadian prog musician whose face is wrapped in white bandages).

Re:Yes, but...

[Cristofori42]

I'd love to see the sheet music for that

Pomp and circumstance

[Cristofori42]

For the love of all that is holy why did they teach it to play that song?! I've been spending all of my years since high school band trying to erase that song from memory after playing it over and over and over for hours on end.

Re:Pomp and circumstance

[sokoban]

pomp (pmp)
n.
1 Dignified or magnificent display; splendor.
2 Vain or ostentatious display. See synonyms at display.

circumstance (sûr'km-stns')
n.
Formal display; ceremony

Seems like the perfect song for unveiling a useless vanity project such as this.

general purpose

[phantomfive]

Are we merely struggling to hard code each human activity as we strive for an all purpose android? Yes, playing the violin is a hard coded activity, but the important advance here is the new dexterity of this robot. It isn't so much an advance in artificial intelligence as it is an advance in mechanics.

Is there a chance artificial intelligence & robotics will ever become generalized enough to make interaction interesting? Absolutely. :) I'll do it myself if no one else gets to it first.

Re:general purpose

[curunir]

Yes, playing the violin is a hard coded activity, but the important advance here is the new dexterity of this robot.

And an incremental one at that...Toyota has had a robot that plays the trumpet for quite some time now. And from what I saw at their corporate headquarters, the trumpet-playing robot is better when compared to human trumpet players than this robot is compared to human violin players.

Though this is no doubt due to the fact that the difficult part of playing the violin is the dexterity whereas the trumpet basically has three on/off finger positions and the rest is in how you blow on it.

Still...the trumpet-playing robot was the better dancer by far.

Obligatory Star Trek reference

[Anonymous Coward]

Remember when Data played the violin for Spock's father in the episode of The Next Generation when he lost control of his emotions due to aging?

Re:Obligatory Star Trek reference

[Romicron]

Actually, my first star trek reference thought was to a different episode of TNG, where Data and his Soong-created "mother" played a duet together, twice. Data noted that his mother/android, in two performances, played the exact same notes in the exact same way, something no human could accomplish. Because of this, he was able to deduce that she was, in fact, not human. Seems a lot more relevant. /// Impressive. I'll be REALLY impressed, however, when they're able to write algorithms complex enough to allow the robot to "improvise", a la jazz.

It's over man; it's over!

[explosivejared]

They can stand back up after being kicked and now can play the violin. Anyone sane could obviously see that this completes their skill set. They'll use the sweet sounding music to lull us all to sleep, and then with their new found balance and agility put the kibosh on us all. I can feel their cold, icy hands around my throat just now! It's over man; it's over!

Re:It's over man; it's over!

[plover]

They'll use the sweet sounding music to lull us all to sleep

and then they'll eat our medicines for fuel!

Oh, man, those things are everywhere! I sure wish I could get insurance against a robot attack! But where?

Re:Robot's sense of time..

[ubergamer1337]

Considering that I'm just wrapping up a semester of violin methods for a music ed degree, I find this achievement more impressive then building a robot to play any other instrument that I can think of because the violin requires extremely precise movements and pressure. The strings take a fair amount to force to depress, but the instrument itself is rather fragile. Also, to get an even sound out of it, the bow pressure has to constantly and smoothly changed while moving.

Pretty awesome, because...

[markov_chain]

...it will obviously lead to better fuel economy and more reliable engines ;) Man, does Toyota have a work-on-your-own-projects day like Google?

Re:Pretty awesome, because...

[JanneM]

...it will obviously lead to better fuel economy and more reliable engines ;) Man, does Toyota have a work-on-your-own-projects day like Google? Toyota is a very big company, with resources invested in a lot of areas that have nothing to do with car design. In addition, they've most likely seen the end of automobiles as a growth market coming and have been actively betting on robotics as the next big-ticket growth area within the next twenty years or so.

Re:Pretty awesome, because...

[Sean0michael]

While I am surprised that they built a violin-playing robot, I am not surprised by their advances in robotics. Since they probably use similar robots in all of their plants to manufacture cars with increasing precision, they would naturally develop the same knowledge to build the robot they did. Notice that the robot's real advances are not in software, but in mechanical precision -- being able to move in the very technical manner of a violinist's bow and fingers.

If their robotic arms can play the violin, what kind of advances in automobile manufacturing could they make with such a huge core skill?

Article asks silly questions...

[Yold]

Specialized robots work better than general-purpose ones (DUH!). Creating a robot that is as capable at general tasks as a human is pointless, at least from the economic standpoint (unless you need a Terminator). Humans are cheaper than robots. Imagine the R&D and production cost involved in creating a robot as agile as the human body. Then, imagine fixing such a robot.

Robots perform special tasks better than humans. Surgery is an obvious application, as the summary pointed out. What could be more steady than a hand with hydraulic (or whatever they use) joints. If something is able to play the Violin, it very well may be able to cut you open along a very precise line, remove a cancer/organ/ while the surgeon is sitting on his butt, operating a computer. Surgery is very tiresome from what I understand (I worked in the dept. of orthopaedics in college), and I'd imagine if this is coupled with the proper software and human interface, it would work splendidly for medical purposes.

I'd think the Medical field would be the most interested in this tech. Surgeons could maybe even perform an extra surgery a day ($$$$$$$), and Hospitals usually have big moolah to spend on fancy-schmancy tech.

Re:Article asks silly questions...

[WK2]

Creating a robot that is as capable at general tasks as a human is pointless, at least from the economic standpoint (unless you need a Terminator). Humans are cheaper than robots.

They are now. Calculators used to be more expensive than hiring ten people to do the job.

Then, imagine fixing such a robot.

Since we're talking about the distant future, I imagine the thing will eventually be able to fix himself. Or be fixed by his peers.

Re:Article asks silly questions...

[E 99]

I'm not saying it's impossible, but surgery is just the type of thing that robots would generally NOT do very well. Every surgery is different. Robots offer an advantage over humans when every repeated action needs to be identical. Surgery requires looking into a space and visually identifying the various structures... all while blood is repeatedly covering them up, and they're jostling around. That's the type of thing that the human brain is exceptionally good at, and robots are ridiculously bad at. Add the fact that there is no margin for mistakes, and I don't think we're going to see robotic surgeons any time soon.

Re:Article asks silly questions...

[veganboyjosh]

Calculators used to be more expensive than hiring ten people to do the job.

Then, imagine fixing such a robot.

Since we're talking about the distant future, I imagine the thing will eventually be able to fix himself. Or be fixed by his peers.

And, eventually the thing will be able to "fix" the 10 humans doing his job less efficiently...

Re:Article asks silly questions...

[Zironic]

the idea is that the robot holds the scapel and the surgeon is remote controlling, not all robots are autonomous.

Re:Article asks silly questions...

[timeOday]

Robots offer an advantage over humans when every repeated action needs to be identical.

That's pretty much like surgery. Most medicine isn't about making up stuff on the spot or putting "your own riff" on a procedure, it's about carrying out a specific procedure in a specific way. Of course it's adapted to the specific variations in each body, but that's not necessarily so hard. A lot of it is like airline pilots. What they do is very easy, but we respect them because it's so important.

Re:Article asks silly questions...

[curunir]

Creating a robot that is as capable at general tasks as a human is pointless, at least from the economic standpoint

Toyota's plant in Toyota City produces one automobile every minute. And they have many more plants around the world. Much of the assembly line is already automated, but there's still a lot of steps that require humans. Any advances that Toyota makes that allow them to replace humans with robots that can do a more accurate/faster job will pay for itself when you consider the increased production capabilities and the decreased defect rate.

Playing the violin is a nice demo, but attaching the door hinge to the car (and the like) will be what makes the R&D worth it.

Re:Article asks silly questions...

[JaredOfEuropa]

Specialized robots work better than general-purpose ones (DUH!). Creating a robot that is as capable at general tasks as a human is pointless, at least from the economic standpoint (unless you need a Terminator).

My Roomba does a good job of vacuuming the floor. However, it does an awful job at loading the dishwasher or changing the cat's litter box. How is that "working better"? General purpose, humanoid robots have many advantages. They can go where humans go, use tools that humans use, but they are stronger and (some day) will be more dextrous and accurate. Also, they don't unionize (well, hopefully...).

In my own household I'd prefer a general purpose robot that loads the dishwasher, cleans and vacuums, does grocery shopping, instead of having a specialised robot for each task. If I need a hole drilled in my wall, I can just ask the robot, and he will be able to use my tools. If my neighbour sprains her ankle, the robot can help her up the stairs. That sort of thing... versatility goes a long way.

Re:Article asks silly questions...

[Grishnakh]

Versatility is expensive. Your Roomba costs about $250; how much is a human-like robot worth? For the cost of one extremely complex but versatile robot, you could have an army of cheap purpose-built robots, each specializing in their own task. Human society has already proven that specialization of work produces more efficient economies.

yeah, but...

[veganboyjosh]

will it stay standing if i kick it?

Re:yeah, but...

[jamesh]

A soft drink machine was invented that would complain if people kicked it or tried to tilt it. So it got kicked and tilted more than any other.

Some cars, when parked, ask people to move away if they get too close, so people deliberately get close and try and taunt it.

A new digital media format is released, with a claim to being uncrackable, so it gets cracked very quickly.

So logically, what happens when a robot gets invented that's sole claim to fame is that it won't fall over, even if kicked?

And now we find that even a robot who's sole purpose is to play the violin is going to get kicked too, just to see what happens...

I think i'll invent a line of robots who's sole purpose is to whack you over the head with a cardboard tube if you kick them or other robots over, or just generally abuse technology for your own amusement. Then i'll release version 2 which features a crowbar instead of a cardboard tube. I'll make a fortune selling them as guards for kick-overable robots, vending machines, cars, and DVD's.

Re:yeah, but...

[JanneM]

I think i'll invent a line of robots who's sole purpose is to whack you over the head with a cardboard tube if you kick them or other robots over, or just generally abuse technology for your own amusement. Then i'll release version 2 which features a crowbar instead of a cardboard tube. I'll make a fortune selling them as guards for kick-overable robots, vending machines, cars, and DVD's. It'll have a hard time protecting anything, being constantly mobbed by people dressed in tight leather straps, handcuffs and hoods with signs reading "Hurt me, I've been bad."

Ah but ....

[icepick72]

..can it play a good dirge on the violin

Re:Ah but ....

[Digestromath]

I'm holding out for the bagpipe playing robot. Put him in a kilt and let the good times roll.

I could actually see one of these robots sitting in the corner of a bar. Someone in a leather jacket walks up, gives it a good bump and it jumps into action. Like a ultra advanced jukebox. Not particularly practical... but what the hell.

Interaction interesting?

[istartedi]

It's hard enough to find *people* who are interesting. Not impossible, mind you; but there are an awful lot of dolts out there. We'll have to *surpass* the quality of humanity before we produce robots that don't fail the "intesting" QC check at unacceptable levels. It's not a total loss though. The failures might be useful as politicians, supermodels, talk-show hosts, morning DJs, etc.

But can it play...

[$criptah]

The Flight of the Bumblebee? Until then, I would rather see Toyota focusing on bringing back trucks like FJ80 instead of the scum infested soccer-mom mobiles they tend to produce now.

Re:But can it play...

[Anonymous Coward]

Can *you* play The Flight of the Bumblebee on a violin? Sure, that robot is not the best violinist of the world, but why does everyone dismiss "being better than the average person"? After all, all you need is to be better than the average person at everything to pass the Turing test.

Re:But can it play...

[$criptah]

Sarcasm does not show well in Slashdot posts... Guilty as charged. I do not like violins that much, but I know several people who could play that pretty well. I heard it live in my living room and it was fun-fucking-tastic. Now, Toyota has spent how much money on this? I bet it cost them a nice penny. Yet I still cannot find a suitable 'yota SUV produced in the 21st century. Don't get me wrong, I believe that teaching a robot how to play anything other than 'play dead' is an awesome achievement. I just wish they Toyota spent more time on designing cars like FJ80.

they should modify it

[RHSC]

to play guitar hero. I'd love to see machine vs program on "Fire and the Flames"

Re:they should modify it

[veganboyjosh]

There's a garage/punk band from Tucson (the weird lovemakers) who wrote a song originally called "John Henry Bonham", about the drummer in band being approached by his bandmates, and being told that he'd been replaced by a drum machine. He challenged the drum machine to a 45 minute song. He kicked the machine's ass, but then he died at the end. Good stuff.

"small" advances

[loonicks]

It seems there have been small -- or maybe even strange, impractical -- advances in robotics

Welcome to the world of research. It takes a lot of work to make small advances like this one. The point of research is to solve specific, difficult problems. I'm willing to bet there were other reasons for this project.

Re:"small" advances

[RHSC]

sure. such dexterity in a robotic arm, combined with machine precision, would increase efficiency in automobile production by an order of magnitude, but as these advances in automation always do, it will likely take away many semi-skilled jobs as well. On the other hand, now is a good time to be a robot technician.

Re:"small" advances

[RHSC]

so you're saying that this doesn't have an obvious application in manufacturing processes? And you're calling me stupid?

Re:"small" advances

[QuantumG]

I'm willing to bet there were other reasons for this project. Gee, ya think. I wonder how a multinational group of companies that grew out of a core competency of factory automation could possibly make use of advancing robot technology.

Re:Violin???

[DustyShadow]

Those were invented years ago man

Microsoft did this.

[WK2]

Is there a chance artificial intelligence & robotics will ever become generalized enough to make interaction interesting?

Microsoft recently had an AI Santa Claus you could talk to over some service of theirs. It was definitely interesting.

Re:Microsoft did this.

[nweis]

It's not the same thing, but you reminded me of this website.

Yeah, but...

[Anonymous Coward]

...can it play the world's saddest song on the world's smallest violin?

I guess I'll say it

[chelsel]

I for one welcome our violin playing overlords

Re:Yes, but...

[shawn443]

Linux playing drums I wonder what OS does drive this though.

Lots of joints.

[Seumas]

The robot has 17 joints in both of its hands That's one robot muthafucka gonna have one hell of a Plaid Pantry munchy trip comin' up.

Some other violin players

[prakslash]

Some other "robotic" violin players:

(a) Violano-virtuoso: Video link Considering it was made almost 100 years ago, isn't bad! This one used discs to rub against the strings to produce the sound. The Toyota robot uses the back and forth motion of a bow which is definitely more complex.

(b) The violobot: Pic and Text link Video Link Sucks!

(c) An attempt at Penn State from 10 years ago in a research project Link. Made mostly noise. Probably abandoned.

Re:Some other violin players

[TerovThePyro]

For something similar, try the House on the Rock outside of Madison, WI. Very interesting and odd place all around. There are several music rooms with set-ups of chamber orchestras, quartets, and band ensembles all done with automated machines. From what I saw there, the machines are very hard to keep in tune, which gives something like the second video. However, there were a few rooms which sounded quite neat, especially given that it was all automated.

Links:
http://www.thehouseontherock.com/HOTR_AttractionMain.htm http://www.roadsideamerica.com/attract/WISPRhouse.html

Re:Some other violin players

[NiteShaed]

The House on the Rock also figured prominently into Neil Gaiman's novel "American Gods" as a place of great mystical power......

Bold research on comprehensive AI has halted

[presidenteloco]

it seems, with perhaps Marvin Minsky as an exception, but we need a new guard.

Everything is understanding the nth degree of optimizing Bayesian network inference,
usually applied to a very specific toy problem.

Nothing wrong with that research. Not really knocking it.

But where is the research on how a generally intelligent system could choose what to
focus its inference-engine attention on. Where is the meta-logic about prioritization
and pruning of "trains of thought" depending on success of search and progress
and urgency of need to know compared to other concurrent topics.
Where are the systems that can posit and explore multiple incrementally variant theories
of some aspect of the world, and figure out which theory-variant is a better model of
past and present observations. Where is the system that can take in lots of different
peoples' writings or sayings about things and synthesize an ontology and figure out
whose beliefs are the most promising (truthwise) and relevant.
Where is the episodic memory?
Where is the emotion-tagging of experiences and important generalizations,
and the emotion-guided prioritized recall?
Where are the short-term memory blackboards?
Where is the "utterance" theory and theories for how to inform and motivate
other intelligent agents into execution of a cooperative plan.
Where is the AI just for the sheer wonder of trying to put several techniques all
together and see what emerges?

Re:Bold research on comprehensive AI has halted

[porpnorber]

Little steps, of course. When I worked in AI, I used to tell people, First we need to solve the artificial stupidity problem. That will let us do the tasks that everyone does routinely. Intelligence is rare, anyway.

But of course, I agree with you completely. Much science has been passed by in the rush to get useful techniques out of statistical methods. Humans are all about grammar extraction and pragmatics; it's only one level of abstraction beyond where we are now, but we are being very slow in going there because the first order systems are well understood and can be trained without interaction. (Cognitive scientists should be made to do structural and informational analyses of videos of people interacting with infants as part of their training. I think it brings the point home faster than the mathematical results from proof theory.)

Unfortunately, current political thinking (and here I grant I draw in black and white) is that research should be put at the service of applications (and education at that of industry), and this clamps us to those first-order solutions. Disruptive technologies need not apply for grants.

Re:Bold research on comprehensive AI has halted

[coldcell]

Where are they? In Science Fiction, and maybeour kids' dreams. Aside from computational limits, until such things as "emotion tagged recollections" become profitable or interesting for a company, it simply won't happen. R&D is prohibitively expensive, the military and the consumer are the only gatekeepers in that world. Unless you appeal to one of them, it ain't happening.

Re:Bold research on comprehensive AI has halted

[QuantumFTL]

Actually it might be a better idea to start by creating a definition of "intelligence" that everyone can agree on.

Re:Bold research on comprehensive AI has halted

[Yvanhoe]

Powers That Be don't read SF. They saw AI R&D as mostly useless because robots don't exist and (as a researcher on the field pointed out to me in 2001) "drones will never be used in real combat operations". Now robots exist, drones are a strategical tool that strategists failed to identify quickly.

Now robots exist, they play violins but are too dumb to identify a target or even their own position in a changing environment. Now our Powers That Be realize that AI R&D may very well be the next big thing. But even today they are limited in their thinking (you will see a lot of such thinking even here on /.) that "robots cannot match human abilities". They will continue to claim that even after a chess grandmaster has been beaten by a computer or after a robot will be able to juggle with 20 items or play violin at 100 notes a second.

Re:Bold research on comprehensive AI has halted

[AceJohnny]

They will continue to claim that even after a chess grandmaster has been beaten by a computer

For some reason, I read that as *eaten*... ...wishful thinking.

Re:Bold research on comprehensive AI has halted

[miro2]

You might be interested in the Novamente project: http://www.novamente.net/engine/

Re:Bold research on comprehensive AI has halted

[oblivionboy]

Marvin Minksy in some ways set us *back* in terms of AI, especially when you consider his hand in the Perceptron

Bad performance

[chickamade]

This is seriously even worse than MIDI. A good violinist can make much rounder/sweeter sound. The arranger who arranged that song for violin didn't do a good job either, because the bowing (slur/detache) is horrible: too much detache. I think my computer MIDI would come out better with a better bowing arrangement.

On the other hand it must be hard to program the robot for all the movement that since if it had been easy they would probably have it play Bach instead.

Damn I was hoping for a robot capable be shown a score and sight-read, actually sight-perform it! I'm a composer BTW.

Re:Bad performance

[RHSC]

Having the robot play the violin is not so much to show off its musical skills as much as it is a proof of concept demonstrating fine control along many points of articulation in order to mimic a complicated series of rhythmic motions

Re:Robot's sense of time..

[mapkinase]

The movement changes in detache are a bit rough. No legato. No vibrato. Relatively slow piece. No changes in position (left hand).

Way to go!

If you do not like your neighbors in the apartment complex, that robot could be a perfect acquisition for home.

Japanese engineering does not stop amazing me.

Re:Microsoft did this. Whoa! Futurama moment!

[jon287]

Microsoft recently had an AI Santa Claus you could talk to over some service of theirs. It was definitely interesting.

You just said Microsoft, AI and Santa Claus in the same sentence... Why am I thinking of Futurama?

Force control?

[TerranFury]

I don't know anything about this particular robot, but methinx you wouldn't be able to do that with an open-loop sequence of scripted movements. There's definitely some sort of feedback going on. Some sort of force control, perhaps? And I wonder if there's some audio feedback as well to keep it in tune?

Self-awareness is great and all, but I don't think it's going to happen. Whereas there are many interesting challenges in lower-level control, and we do seem to get results there when we work at it.

Re:Yes, but...

[ls671]

Nope it runs on a toy-OS

or was it a violin-playing robot?

[clarkn0va]

The violin wasn't actually playing robot, was it?

db

should be implemented in pixels

[grikdog]

Robots should be graphic actors which mediate between people and machines, such as allowing an autopilot screen personality to parallel park a Toyota. Can you imagine phoning Angelina Jolie (the Grendel's mom version) and telling her you'll be late getting home? Then, in 5 us or less, she tells your kitchen that dinner will be late by 2 hours, so don't defrost the TV dinners yet? There's a crude version of this in Red Dwarf (Holly), a parody of it in M.I.B. (the POS Ford autopilot), and a fully-formed sketch of it in GITS (cybernautic tachikomas). Prior mention goes to Arthur C. Clarke, for the "monitor" (IIRC) in Against the Fall of Night.

Artificial intelligence.

[onefriedrice]

> It seems there have been small -- or maybe even strange, impractical -- advances in robotics

We can draw a picture of a tree and otherwise produce accurate models of our world in many cases, but God's greatest creation will not so easily be emulated. It will yet be a while until we will truly understand how the brain works and how intelligence can be programmed, but it may happen after lots of research. How incredible. It's interesting how people can be fooled into believing such things can be "randomed" into existence. They believe God to be improbable while they do not stop to consider the improbability of what they really believe. Interesting.

Re:Artificial intelligence.

[minasoko]

Off-topic, but improbable != impossible. There is evidence to support the theories surrounding how the human brain could have developed. There is no evidence to support the existence of an all-powerful, omnipotent being that created the Universe. So, I guess for some it's as simple as, improbable plus some plausible theories > improbable plus zero evidence.

Breaking A Million

[ryanisflyboy]

So is it cheating if I use one of these to break a million points on Guitar Hero?

After that I'm going to put him to work on Heroin Hero... I WILL CATCH THAT DRAGON!

Re:Robot's sense of time..

[UbuntuDupe]

In an attempt to prop up my own achievements (I played violin for six years), I agree :-)

But what surprised me about the video was that, while the robot's playing was messy, it appeared to make the same errors and imprecisions that new human violin players make. I don't know if I'd be able to distinguish its playing from a seven-year-old's recital if I had to judge by ear alone.

"Strange, Impractical Advances"

[Invidious]

It seems there have been small -- or maybe even strange, impractical -- advances in robotics repeatedly with demonstrations of robots performing a specialized task. Are we merely struggling to hard code each human activity as we strive for an all purpose android? Is there a chance artificial intelligence & robotics will ever become generalized enough to make interaction interesting?"

Most of the times, these small, strange, impractical-on-their-own tasks are not the result of building to that specific task (we need to develop a robot that plays the violin!) but of developing or bettering a technology that's showcased by that task (playing a violin requires relatively high-speed dexterity, timing, and precision, which is something that would, of course, be great to have in a humanoid robot.)

That being said, there isn't going to be a true general-purpose robot until we either a: create an AI sufficiently powerful that it can learn any unprogrammed tasks the same way a human would, through observing others and experimentation, or b: program all those tasks.

Re:Violin???

[bronney]

i am sorry but she could barely play my piccolo.

Re:mechanically impressive, but...

[bronney]

The question is, why aren't they doing more to make robots self aware?

The real question is, what makes US self aware? We don't even know that, let alone programming it.

Learning Robots

[Deliveranc3]

The thing about teaching a robot to learn is you need a metric for success and failure.

Human's don't even know what ours is yet! (Though we have theories ).

Assuming we want the robots purpose to be making people happy, we haven't even found a way to qualify happiness yet, let alone quantify it.

Psychology is a quagmire, people are diffrent, and we'll need to come from both directions (Psychology and adaptive A.I.) to develop useful heuristic models for A.I.

Re:Robot's sense of time..

[RealGrouchy]

I don't know if I'd be able to distinguish its playing from a seven-year-old's recital if I had to judge by ear alone. The recording of the seven-year-old's recital has the kid's grandmother in the background saying "isn't that sweet?", and the kid's father grumbling that he got dragged to this thing when he could be watching the game.

- RG>

Re:Yes, but...

[rts008]

Yes, but only on a Beowolf cluster serving e-mail to old people in Korea, while in Soviet Russia it is playing you while it is being confirmed by Netcraft, and....I can't go on, someone else needs to either take over or shot me between the eyes!

Me:"Why yes ociffer, I HAVE been drinking while posting, but I had a good reason!"

Officer:"Sir, you are obviously drunk. Did you not have someone you could call to come get you?"

Me:"Well it's like this: i couldn't walk this drunk, so I tried driving- but my house key broke off in the ignition switch. I tried hotwiring it, but caught the trunk on fire-must have been my stereo wiring I did last year-but anyway, I pissed on the fire, and after another liter of beer, I finally got it out. What? Yes, my penis...Yes, I finally got it out. What? Oh, yes it did take another sixpack to put the fire out once I did get it out of my pants- that helped a lot, I can tell you!
Oh yes, so I was too drunk to walk OR drive by then you see, so I grab this keyboard in the bar....What?
Well damn right I went back into the bar! Haven't you been listening?
1. Too drunk to walk, so I have to drive.
2. Can't drive.
3. What else to do but go back into the bar?...I ain't stoopid!"

Officer:"Where was the computer (that you grabbed the keyboard for) located? Was it a public accessible computer?"

Me:"Wha??? Uhmm, remember, I was too drunk to walk....."*gets cuffed and tossed into the trunk of the POlice van-have you ever really thought about how 'small' that donut/spare-tire actually is up CLOSE!- I thought not!*[apply sarcasm and Monty Python-type filters here @heavily +3]

With your UID, when you make this type of reply, you might want to get creative.
But then again, it could just be my Blood Alcohol Content posting!

Nanotech

[Weaselmancer]

Nanotech needs to get involved somehow. That way you could guarantee that what you're hearing is the world's smallest violin, and it's playing just for you.

Innovation fueled by xenophobia

[HungWeiLo]

Japan, as well as many other industrialized nations, will face a dire shortage of workers as their population ages rapidly amidst a very low birthrate. While North America, and to some extent, Europe, have paths laid out to take care of this problem partially with immigration, Japan is not at all immigrant-friendly. Immigration to Japan is very difficult. Crime is universally blamed on ethnic Koreans and Chinese, despite the fact that they commit less than 1% of all crimes in Japan. Other than Korea, Japan is probably one of the more racially homogeneous societies in the world.

What this means is that the Japanese would very eagerly pour their resources into creating robots which can man factory floors, clean houses, and cook food. Expect lots of jaw-dropping innovation in robotics from Japan in the coming decade, simply due to pure demand.

Toyota plays fiddle

[tronkel]

Ah yes, I believe they will be using this system in the forthcoming upgrade to the Toyota Prius model announced at the Frankfurt show. It is to be called the "John Doe" series. It automatically switches over to "auto-violin DRM-free mode" mode when within RIAA hearing range. RIAA can't prosecute a robot -it's got no money!

"myopic" mod needed.

[doug141]

And 640K ought to be enough for any program, and a computer will never beat a man at chess...

By 2050, a robot will read what you wrote, download violin skills and compose and play the most beautiful piece ever, better than any man, just because it can.

Specific tasks / generality

[dyftm]

"Are we merely struggling to hard code each human activity as we strive for an all purpose android? "

No. While working towards this specific task, I'm sure they will have solved problems that exist more generally in AI/robotics. It's like when Fermat's last theorem was proved - the fact that it was proved itself was relatively insignificant, the problems solved and the maths generated along the way were of huge value to mathematics.

Was on...

[Tastecicles]

...the BBC this morning. I thought to myself, "Well cool, now all we need is to teach it to ride a unicycle and juggle and our cybernetic slave shortage will be solved!"

Geeks still have a chance...

[teddygraham]

"Is there a chance artificial intelligence & robotics will ever become generalized enough to make interaction interesting?" Who needs a girlfriend when you have robots that can perform 'interesting interactions'?

Software development

[hernyo]

Goddamn, our so-called "state-of-the-art" is NOTHING compared to this. Desktop software development is fucking 20 years behind what these guys are doing.

Maybe I should just quit.

Robot nerds

[El Yanqui]

We teach robots how to play chess and the violin. No wonder they're getting kicked.

No fate ...

[KillaBeave]

... but what we make?

::Tightens Chinstrap::

Re:Yes, but...

[EMeta]

No, but it'll keep up if you hum a few bytes.

Dishes and Laundry Please

[darrint]

I don't care if it interacts in "interesting" ways. Do my housework. Also, mow the lawn.

New Polls indicate...

[Fysiks Wurks]

Toyota Robot edges out John Edwards and is in a dead heat with Clinton and Obama with potential Iowa Caucus voters.

Another poll indicates 100% support from electronic voting machines in all states.

Jesse Sullivan

[Thelasko]

The dexterity of those fingers are very impressive. Is each arm is purpose built? One arm built to hold the bow and one to press the strings. If each arm is identical I would like to see it applied to the field of cybernetics. Right now researchers can obtain enough signals to control most of the movements of an arm but don't have an arm capable of making those motions. Someone like Jesse Sullivan could benefit greatly, even if the arm was only used in a laboratory environment, because I suspect it's quite heavy.

Next special purpose robot...

[Organic Brain Damage]

...with 17 joints in the hand, can it be long before we have special purpose hand-job robots?

Toyota succeeds where House On The Rock failed

[dpbsmith]

The quirky but quite marvelous House on the Rock, a tourist attraction in Wisconsin with geek appeal, is a Barnum-like pseudo-museum that is genuinely awesome, crammed through of things many of which are not quite what they seem. You can't quite call them fakes because the management no longer makes explicit claims of much of anything. Among the marvels are numerous room-sized assemblages of automated musical instruments, automated the good old fashioned way a la player pianos and "orchestrions."

There is a suggestion that these are historical Victorian contraptions, but in fact most of them were built in by showman/architect Alex Jordan and his colleagues.

Quite a few of them feature stringed instruments, and all of them have been devised to look the way you would hope a mechanical violin-player would look: clockwork gadgetry draws a traditional bow visibly across strings. But while visually effective, they are fake; they are silent, violin-like notes being produced elsewhere from an organ or synthesizer.

The collection does include one genuine coin-operated violin playing machine, a "Violano-Virtuoso," but instead of a bow, it uses motorized rosin-covered wheels. Mechanical fingers press the strings down, producing a correct pitch but no vibrato. The sound is just as harsh, unmusical, and "mechanical" sounding as you'd expect. Unlike player pianos and carousel orchestrions, where the mechanical sound can be perceived as having a charm of its own, the Violano-Virtuoso just sounds unpleasant. Apparently, it is really difficult to get a pleasing sound by bowing a string.

Toyota's demonstration thus has added piquancy from the fact that not only is the achievement impressive, but it actually is something that people had been trying to do for about a century, without success.

I believe this is a direct quote

[Lordnerdzrool]

From Yo-Yo Ma after the cello playing edition was revealed:

"DEY TOOK OUR JEBS!!"

anthopomorphic wrestling robot

[peter303]

PBS Wired did a piece of wrestling robots in Japan (GeekDad segment). These robots have arms and legs, stand, summersault, rise from falls. Its amazing what they do now.

Stylish hardware, what's the control system like?

[Tom Womack]

This sounds like a nice engineering exercise - very small motors with adequate torque, very small angle encoders, elaborate chassis to contain it all with the right number of degrees of freedom. The interesting part is always the control software, and I can't work out from the article whether the robot is running software that can play the violin ab-initio, or playing back a sequence of moves figured out by infinite grad-student labour.

I'd probably guess the latter - 'can it play any other piece' and 'can it play _this_ violin' (offering one of a slightly different size - with the popularity of the Suzuki method, children's violins are made in a variety of sizes) would be the killer questions.

I've heard of robots (WABOT-2 back in 1985) which can play piano from a score - OCR on musical notation turns out not to be especially difficult - but the extra mechanical linkages in a piano mean that you can press the keys in almost any way and get something note-like.

I'm rather more impressed that Toyota have managed (http://www.toyota.co.jp/en/special/robot/) to get robotic lips to play the trumpet - that's a nasty problem in engineering active soft materials.

Re:Robot's sense of time..

[djh101010]

What I find exasperating about this is that they spend thousands, millions on getting this robot to play the violin, then they neglect to code the song correctly. The robot may be actually doing a pretty good job, it just doesn't sound musical! It's like spending millions to train an athlete then only letting him do Morris dancing.
Yup. Cheap guitars are a perfect example. Not a one of 'em is playable, and it's ALWAYS for the same reason - the bridge is in the wrong place so the intonation is horrible. Yet, the effort make a cheap guitar with the bridge in the right place, is exactly the same as it is to do it poorly. So you end up with a kid who wanted to learn guitar, with something that has the strings too high so it hurts, and the bridge too close so it gets sharper the higher up the neck they go. Not so much a guitar, as a "guitar-shaped-object". Hardly a way to encourage a beginning player. Like your song chips you mention, this is a case of something that's being made to be sold, rather than to be used. (And, it's probably got lead paint on it...)

As I have no mod points,

[Facetious]

thank you. I'm still giggling.

Shirow Masamune / robotic control software

[Karl the Pagan]

Perhaps as in the "Ghost in the Shell" fiction series each function of a robot or cyborg's abilities will be linked to specialized software tailor made for that application. The power users will combine and tweak their control software, hackers could inject viruses, or feedback in cybernetic systems could drive users mad.

It's all been predicted... now we just need to show that the alternative (AI) is possible and test if it is more desirable.

Re:Robot's sense of time..

[Creepy]

Heh - and my first thought was "don't quit your day job."

Yoda would probably say "technical achievement not master of violin make," but since he died long ago and far, far away, I should probably not think of it in Yoda-isms and stick to "wow, that's pretty impressive for a robot."

Still, I've seen dexterous robots and am a decent cellist, so I know how far it has to go. In some ways it reminds me a bit of listening to an orchestral piece in MIDI - all the parts are there and the piece itself may be amazing (no, I don't mean pomp and circumstance is), but it lacks the dynamics, articulation and phrasing that make music an art.

Life vs. Robots

[inKubus]

Human-developed robots will continually approach the abilities of actual living things but will never acheive it.

God almighty

[Impy the Impiuos Imp]

> Toyota has unveiled a robot that can play the violin.

Fools! Don't they know this kind of R&D will never lead to anything profitable? The correct way to make profit is to loan your money out to masses of US yokels at outrageous interest rates, then have hirelings in poor countries ride herd over automated machinery to call into the US to harass the citizens.

Ouch ...

[jopet]

Making noises with a violin is not necessarily "playing the violin".

Re:Robot's sense of time..

[blitzkrieg3]

The movement changes in detache are a bit rough. No legato. No vibrato. Relatively slow piece. No changes in position (left hand).

Actually, the robot did have vibrato, though not in the typical way. The strings are probably higher than on a typical violin, and the robot will press up and down on the string to "simulate" vibrato. It has the same effect, and I was wondering how it got the sound until I saw in the closeup of the left hand.

Re:Yes, but...

[Inakizombie]

I don't think robots will take that right from basement dwelling nerds any time soon. :)

Re:Robot's sense of time..

[mapkinase]

I missed that in the video. I guess I have to see it again.

Quality...

[TheEgolf]

Sure you can listen to a MID/MIDI of a symphony from Mozart, and it'll never sound like an actual live concert band playing it, just like a robot won't sound like an actual concert band.

Suzuki method?

[MMHere]

Does it learn to play via the Suzuki method. If so, does Toyota have to pay royalties to or cross-license patents from Suzuki Company, one of their competitors in Japan?

Robotic fingers.

[PrefersVMS]

To hear of an achievement such as this, is a sign of hope.

Not for the robots, but for the amputees who need better hands to replace what they've lost.

If you look only at the surface, would one care what is happening in the robotic world?
Probably not. But...if one were to see what such an advancement could do for those of
us who have lost limbs? Then this is an amazing achievement.

And I'm speaking from personal experience. For 56 years, I walked on two legs. Eight
years ago, I tested positive for type 2 diabetes. Last January, I lost my right foot,
from above the knee, to this nasty disease. Now, I live bound to a wheelchair. I can't
get any work, because the right leg is missing---how does one drive without the right
foot? Yes, I know there are controls to help with that--first I have to be evaluated to
determine which set of controls I can use, then buy the controls and have them installed
on a vehicle. Not just any vehicle, but more than likely, a van or minivan. The cost?
Now there's the limiting factor. The cost for the evaluation and training to drive
is over $1000USD. The controls' price ranges from $250 to $2000USD. And then there's the
cost of the vehicle. For example, you, as a two-legged individual could purchase a new
Honda Odyssey for about $35,000. Fully equipped. For a similar model, for a right-leg amputee? Add another $15,000. And I'm sorry to say, that insurance does not cover this.

Sorry to sound like I'm ranting. It's just personal. So, yes, I'm praying that the work in robotics can be transferred over to the amputees. And that would be the real miracle.

Re:Robot's sense of time..

[HeadlessNotAHorseman]

In an attempt to prop up my own achievements (I played violin for six years)

Boy, your arms must be tired! Ba-dum-ching!