Triumph of the Cyborg Composer

I thought i'd have a try at getting Buzz to do a similar thing to the first piece on the aforementioned article, after about an hour I got this -

Cool Awepittance -- I did a GOogle search of WATMM to make sure I wasn't jazzbanding, but didn't see your thread. That's awesome you got to sit in on a class!

I really want to know more about how his latest program works -- how it 'learns' what kind of thing he wants. I also want to know exactly how he went about 'taking apart Bach's works,' as the article describes it, and 'putting hte pieces together again' -- I wonder what specifically that means? That is the fascinating thing. Because for me, a lot of the joy of electornic music is always the 'WAT IS DIS' / 'how'd they do that?' question. And for this guy to have that level of understanding of what common tropes or patterns make up a particular composer's style, I think, justifies his 'copied' versions as creative and genius -- because it's not just accident or something that happened to come out right. He spent time doing ... whatever he did ... in order to really understand the musical principles behind the works. I would love to hear him give a lecture going through a bunch of composer's works and pointing out all the similarities of the chord progressions, high and low notes, etc. -- I guess that's what musicologists do, but this just sounds so much more detailed.

I will also definitely seek out those Chopin clones; that sounds awesome!

*puts on Invisible Touch*

  On 2/25/2010 at 2:41 AM, encey said:

I also want to know exactly how he went about 'taking apart Bach's works,' as the article describes it, and 'putting hte pieces together again' -- I wonder what specifically that means?

Simplistically speaking, the program probably identified phrases in the original bach works and altered the scale or mode of the phrases and recombined them. If you listen to the new bach invention he's put up, the beginning phrase is copied from Invention 8.

  On 2/24/2010 at 10:29 PM, BCM said:

computer can find patterns in groups of notes != computer will ever be able to feel things the way a living organism does

i have to disagree here, sir. It has been mathematically proved that everything in this universe can be reproduced by a turing machine. So one day, computer may be able to feel the same way we does (all the more so if you're trying to reproduce the way a bee feels).

Also lol at the, "this computer makes neat pieces of music but it has no soul" argument. I wouldn't be able to produce music this good. Does that mean i have no soul ? Anyway, this exact same argument is used when it comes to chinese interpreters outperfoming occidental ones.

  On 2/25/2010 at 10:04 AM, Babar said:

Also lol at the, "this computer makes neat pieces of music but it has no soul" argument. I wouldn't be able to produce music this good. Does that mean i have no soul ?

oh jesus lol, thanks for bringing a smile to my face

you're welcome. Can someone post a track from http://artsites.ucsc.edu/faculty/cope/mp3page.htm. I can't reach the track, too many users. I'd like to hear the mazurka. thanks.

ditto babar. specifically the chopin one....

  On 2/25/2010 at 10:04 AM, Babar said:

  On 2/24/2010 at 10:29 PM, BCM said:

computer can find patterns in groups of notes != computer will ever be able to feel things the way a living organism does

 

i have to disagree here, sir. It has been mathematically proved that everything in this universe can be reproduced by a turing machine. So one day, computer may be able to feel the same way we does (all the more so if you're trying to reproduce the way a bee feels).

 

one day i might poop butterflies and they wouldn't even smell bad.

you're missing the point. Creativity is not about analytical imitating of natural phenomena. in this case here, Cope is the composer, not his software. It is his creativity that makes these compositions sound good, he spent many years manually tweaking algorithms which do just that. in the end it's the same as if he wrote each composition with pen and paper, only he built himself a far more complex pen and paper.

now a software that would actually be creative, that would experiment with its own algorithms, explore its own boundaries only to come up with ways to surpass them to create new, quality compositions or images or anything, question itself what it is doing etc…. in fact you'd need software which you programmed for dealing with your bank account but it would one day say: "look, Dave, i wasn't in the mood to do your bank account today so instead i spent the whole day reflecting on this bunch of pixels i pooped out." or something.

it would be so utterly complex. The problem with computational problems is that you have to know every single step of the process and be able to exactly define the problem in order to make an algorithm from it. A computer is good at one thing, one thing alone: precisely following a set of mathematical instructions, fast.

it's so easy to say, oh well, a bee is not that complicated, one day we'll be able to code the feelings of a bee. go on, define how a bee feels.

  On 2/25/2010 at 3:02 AM, Rabid said:

  On 2/25/2010 at 2:41 AM, encey said:

I also want to know exactly how he went about 'taking apart Bach's works,' as the article describes it, and 'putting hte pieces together again' -- I wonder what specifically that means?

Simplistically speaking, the program probably identified phrases in the original bach works and altered the scale or mode of the phrases and recombined them. If you listen to the new bach invention he's put up, the beginning phrase is copied from Invention 8.

cheers

*jots it down*

  On 2/25/2010 at 10:04 AM, Babar said:

  On 2/24/2010 at 10:29 PM, A/D said:

computer can find patterns in groups of notes != computer will ever be able to feel things the way a living organism does

 

i have to disagree here, sir. It has been mathematically proved that everything in this universe can be reproduced by a turing machine. So one day, computer may be able to feel the same way we does (all the more so if you're trying to reproduce the way a bee feels).

once again I disagree! Although I'd love to see this mathematical proof. However, what I mean is: the difference between (a) to display an acceptable mimicry of human emotion and (b) to feel the way a human feels is unknowable (what do computers feel?) and vast. I do not mean we'll never know how other organisms or machines feel, but as of now, the mechanism for gathering that information is outside of our imagination.

  On 2/25/2010 at 10:04 AM, Babar said:

Also lol at the, "this computer makes neat pieces of music but it has no soul" argument. I wouldn't be able to produce music this good. Does that mean i have no soul ? Anyway, this exact same argument is used when it comes to chinese interpreters outperfoming occidental ones.

yes lol I totally agree

once you realize our feelings are a tool, an illusion that helps us cope with complex problems and situations, we'll understand that there is nothing overtly special about them. simply a byproduct of an overactive brain learning how to survive.

of course they aren't simple at all, but at some point in the future we will be able to simulate emotions just the same way our brain does.

did Santa Clause tell you that, on your Holo deck?

why would a computer need feelings?

it should only understand human feelings in some sense to be able to reproduce them.

this thread is just reproducing the debate that was avoided in the article: a romanticised notion of creativity vs. a "cold" one.

cope believes the way humans compose is no much different than how his creations do: copying something else.

what happens if emily howell interacts with another emily howell? if the results of this interaction are stored in some database that ellie can use, then the emilys use ellie's composition as another starting point and start composing. add some decision making with an element of randommness and you have instant creativity.

  On 2/25/2010 at 2:01 PM, phling said:

  On 2/25/2010 at 10:04 AM, Babar said:

  On 2/24/2010 at 10:29 PM, BCM said:

computer can find patterns in groups of notes != computer will ever be able to feel things the way a living organism does

 

i have to disagree here, sir. It has been mathematically proved that everything in this universe can be reproduced by a turing machine. So one day, computer may be able to feel the same way we does (all the more so if you're trying to reproduce the way a bee feels).

 

precisely following a set of mathematical logical(this word brings less confusion) instructions, fast.

Maybe in a few decade we would be able to map a man's brain into a computer. We would map every single synapse and neurone of his brain and feed them to the IBM simulator (remember this cat brain thread we had a couple months ago ?). What then ? Would you say it's just a pale copy of the man's mind, just because it follows logical instructions ?

Anyway, brains rely on neurons, which rely on physics, i.e. maths applied to reality : everything follows logical rules.

To me creativity deals with what is new. But I think there are different levels of

1. if this program can make new songs in the style of beethoven, then it is creative (maybe more than eletronic musicians remixings tracks...)

2. if it can copy a musical style but is not able to create a new one then it's not creative.

3. same goes for genres.

About 2 : well i bet if you take 10 beethovs tracks and one zillion tracks made by this program and feed them to it, it will still make the same kind of stuff, because it rigorously follows beethov's style. That's stil pretty cool.

About 3 : almost no one on earth is able to create a new musical genre.

Also, what cope implies is that music works like genetic material : it's a big meme. When you're making music, you're not creating something from scratch, you're mutating it from what you've already listened to. Maybe we should try to lock a baby away in a cellar, and patiently observe him through subtly hidden cameras. Interacting with him is not allowed. i bet he will never sing a song.

  On 2/24/2010 at 10:29 PM, A/D said:

once again I disagree! Although I'd love to see this mathematical proof. However, what I mean is: the difference between (a) to display an acceptable mimicry of human emotion and (b) to feel the way a human feels is unknowable (what do computers feel?) and vast. I do not mean we'll never know how other organisms or machines feel, but as of now, the mechanism for gathering that information is outside of our imagination.

I was refering to the church-turing thesis. It states that everything a mechanic machine can perform can be done by a turing machine (i.e. a computer). But maybe i'm mixing things up by implying a brain is a mechanic machine.

but i found this :

  Quote

Alan Turing demonstrated, in his 1950 paper (Mind 59:433-460), with calculations, the infeasibility of cognitive machines when explicit programming was their only knowledge acquisition tool (cognition could be achieved only with the addition of an interpreter and humanized interfaces). The authors show that Turing’s main principles, the addition of an interpreter and humanized interfaces, may be replaced by sequential algorithmic programming when the modalities of receptors are taken into consideration. These modalities lead to a fundamental law in biology - the law of specific nerve energy - that relates consciousness to explicit neuronal activity (Neuronal Correlate of Modality). This law may be used to prove that “conscious” machines can exist, and can exhibit forms of consciousness similar to human consciousness. The design of such a conscious machine, a tactile-visual humanoid robotic machine, has already been implemented

  On 2/25/2010 at 10:04 AM, Babar said:

Also lol at the, "this computer makes neat pieces of music but it has no soul" argument. I wouldn't be able to produce music this good. Does that mean i have no soul ? Anyway, this exact same argument is used when it comes to chinese interpreters outperfoming occidental ones.

yes lol I totally agree

and i disagree. Everyone knows chinese have no soul. it was irony.

  On 2/25/2010 at 6:53 PM, phling said:

did Santa Clause tell you that, on your Holo deck?

Are you implying god insufflated a soul into adam ?

  On 2/25/2010 at 6:48 PM, theSun said:

once you realize our feelings are a tool, an illusion that helps us cope with complex problems and situations, we'll understand that there is nothing overtly special about them. simply a byproduct of an overactive brain learning how to survive.

wow, that's pretty bold. and contradictory. how could an illusion be a tool? if they're so useful, why are they not special? sometimes emotions are very useful in our survival; other times they lead to our suicide. i don't see how you came to this conclusion.

we may never see eye to eye on this, but i don't think you have this too well thought out.

why can't an illusion be a tool? maybe i shouldn't have said "special." i meant that they are not unique to humans and they are entirely based upon chemistry and electricity. of course our emotions are more sophisticated, but we have the means (human brain) to explore those emotions through communication, creating a common perception of them.

the usefulness of emotions is applicable only in the evolutionary setting. we did not evolve to live in this technological world, so our emotions have strange consequences at times, especially when you get into the whole "what is mental illness?" debate. as we try to categorize things (emotions) that are far more complex than the systems (modern psychology) in which they are categorized, it is no longer possible to derive useful information from such things. we must look at emotions in the environment that mankind grew up in. how can a feeling of vengeance, love, creativity or lust be useful such that the bearer of said feelings can fuck more bitchez?

feelings help us analyze situations in ways logic would never predict, this diversity in thought is one of the reasons humans have been able to survive and adapt to new environments and become the first species on earth to use tools extensively, and rely on them completely.

i have it pretty well thought out, i have a hard time trying to summarize though.

  On 2/25/2010 at 8:20 PM, theSun said:

(a) the usefulness of emotions is applicable only in the evolutionary setting.

 

(b) we did not evolve to live in this technological world,

 

© so our emotions have strange consequences at times,

 

(d) especially when you get into the whole "what is mental illness?" debate.

I don't quite understand just how (d) is related to © or what © has to do with (b), so that sentence is hard for me to follow.

I don't understand what you mean by (b) -- I take it you will agree that we do live in a technological world, and that we are a product of the evolution of our species. So in that sense, we did evolve to live in this technological world. But then you must mean something like 'we were not supposed to evolve to live in this technological world.' But as I understand it, the criterion for whether evolution has 'happened as it is supposed to' is whether a species is able to adapt to its environment and survive. So far, we have. So that is not enough reason to conclude that we aren't supposed to live in the world we do. Also, I don't understand evolution as having a specific, already determined 'end point' that a species is 'supposed to' reach. So I don't fully get your claim.

I also disagree with (a). I can make use of my emotion of anger by expressing it to someone in order to get them to acknowledge that they have wronged me, that they were out of line, being an asshole, and should apologize or do whatever is required to correct for that. This has nothing to do with evolution; it has to do with, I dunno, justice, or respect, or fairness in my relationships with other people. It's a personal and social phenomenon, not a biological phenomenon. So I don't agree that the concept of emotions is only usefully applied in making judgments about evolutionary success or 'fit' or whatever.

All that is just to see if you can further clarify what you're claiming.

  On 2/25/2010 at 7:46 PM, Babar said:

  On 2/25/2010 at 6:53 PM, phling said:

did Santa Clause tell you that, on your Holo deck?

Are you implying god insufflated a soul into adam ?

not at all. Just being realistic here. Sure, you can dream about really smart AI all day, nothing's stopping you. It's science fiction nonetheless.

Examples where AI really works are all very specific solutions, doing one complex task quite well. Google Translate for example i would say is a fairly advanced AI. However it is translating from one language to another without any "clue" of what the text actually means.

Science is far, far, far, faaaaaaaaaar away from simulating the functionality of a cat brain. It was an esoteric catchy phrase and the internet press jumped on it.

I'm also not against artificial creativity, in fact i've dabbled around with genetic algorithms and neural nets in an art context myself with mixed results. From experience i can tell you, these things work well with finding a solution to a clearly defined problem in an isolated environment. An aesthetic sense however, which in my opinion would be absolutely necessary for a truly creative AI, is a complex high level problem we simply don't understand.

read this book imo: http://www.amazon.com/Evolutionary-Computer-Music-ebook/dp/B00192YJUK/ref=sr_1_4?ie=UTF8&s=digital-text&qid=1267129676&sr=8-4

The computer cannot understand the music.

It's all really cool but it's just doing what it's programmed to do, by man.

  On 2/25/2010 at 9:19 PM, encey said:

  On 2/25/2010 at 8:20 PM, theSun said:

(a) the usefulness of emotions is applicable only in the evolutionary setting.

 

(b) we did not evolve to live in this technological world,

 

© so our emotions have strange consequences at times,

 

(d) especially when you get into the whole "what is mental illness?" debate.

I don't quite understand just how (d) is related to © or what © has to do with (b), so that sentence is hard for me to follow.

 

I don't understand what you mean by (b) -- I take it you will agree that we do live in a technological world, and that we are a product of the evolution of our species. So in that sense, we did evolve to live in this technological world. But then you must mean something like 'we were not supposed to evolve to live in this technological world.' But as I understand it, the criterion for whether evolution has 'happened as it is supposed to' is whether a species is able to adapt to its environment and survive. So far, we have. So that is not enough reason to conclude that we aren't supposed to live in the world we do. Also, I don't understand evolution as having a specific, already determined 'end point' that a species is 'supposed to' reach. So I don't fully get your claim.

 

I also disagree with (a). I can make use of my emotion of anger by expressing it to someone in order to get them to acknowledge that they have wronged me, that they were out of line, being an asshole, and should apologize or do whatever is required to correct for that. This has nothing to do with evolution; it has to do with, I dunno, justice, or respect, or fairness in my relationships with other people. It's a personal and social phenomenon, not a biological phenomenon. So I don't agree that the concept of emotions is only usefully applied in making judgments about evolutionary success or 'fit' or whatever.

 

 

All that is just to see if you can further clarify what you're claiming.

what he means is that feelings are consequence of the primitive stages of human evolution but are often counterproductive in this modern world, and that what our feelings tell us often clashes with what our knowledge tell us, producing unexpected consequences and often tragic.

you conception of evolution is somewhat wrong. old, useless treats aren't necessarily discarded by the evolution process. also it should be pretty clear that humans have managed to bypass some of the classic dynamics of evolution. human society (and knowledge) changes many times faster than any biological evolutionary process. this means our environment changes much faster than how we can biologically adapt to these changes.

one can clearly see the connection of feelings with evolution. for instance fear and love are strongly correlated to survival and reproduction. I think all feelings are mixtures of some basic ones although i can't say what those basic ones are.

well... for the first 99% of human evolution, our genes were under a barrage of natural selection that had nothing to do with any sort of technology. only within the past several ten thousand years have we begun to develop (through creativity, which i would simply define as problem solving) technology and use it to advance as a species. of course we live in a technological society right now, but this was not true for 99% of the path our genes took to get here, therefore the effect of technology on our individual inner workings is insignificant compared to the effect of common threats/benefits over the pre-civilization humans. that's mostly where our emotions come from; fear, love, hate were all developed unwittingly by our ancestors to give future generations a better chance at survival. at the same time, our social order was becoming increasingly complex. we developed medicines, researched ourselves, etc. we began to make inferences as to what our bodies, the earth and space were doing, and make categories of things (hard/soft, bright/dim, good/bad, happy/sad) as we understand them subjectively, so that we might better understand our environment.

in early medicinal psychology, they began to treat patients with "illness" whether they are too energetic or too violent or too timid. they were all basing their claims off of research and studies that could not be validated simply because we did not have the means to know what was happening in the brain. (we also can never define parameters in which human behavior is "normal", that's another discussion). we were just guessing why person x has ocd about washing his sister 3x a day, and prescribed medication z to treat it. of course the medication has a ton of unintended consequences because humans simply did not know what they were dealing with. this is still true today, we cannot know exactly why person x does strange thing y, but we are much closer. take one list at the side effects of any prescription and it is clear we still have a weak grasp on how our body/brain works as a whole.

of course i'm not suggesting any sort of predestination or "end point" of evolution, dunno how you got that. we're not "supposed" to do anything. the world is the same whether you sleep in a pile of shit or are an NBA player. and of course human emotions are useful to you, they are such a large part of our brain that we have adapted them to our own society. i'm just saying that when you try to analyze them in modern terms, our own social structure blurs the facts. what does "justice" mean to a computer (getting back to the OP)? too complex, too abstract to define. also people of different cultures with different ideas of the self are bound to give completely unpredictable answers when asked what "justice" means. we don't all have the same perception of "justice" because it is so new to us, and a combination of common emotions based on circumstance. but something like fear, happiness or excitement we can all relate to, it's in our genes and is a byproduct of our evolution.

also, every personal and social phenomena depends explicitly on the biological and chemical reactions required by the people experiencing said phenomena.

hope this clears some things up, i tend to be a little ambiguous because i don't often write about this shit.

well said gordo! you summed up what i couldn't

also, i just want to add that humans have been in the making for millions and millions of years and electronic computers have been around less than a century (though if humans had millions of years to make a computer i bet we'd make some pretty sick shit)

Did you and Gordo read the same textbook or something? This seems to be mostly non sequitors and broad statements about some controversial and complex stuff, but you guys understand it. Maybe there's further reading you can suggest? Comments below if you care to elaborate.

  On 2/25/2010 at 10:08 PM, theSun said:

fear, love, hate were all developed unwittingly by our ancestors to give future generations a better chance at survival.

I don't get this. You're saying emotions were consciously created in order to give an evolutionary advantage? On the one hand, emotions do not share an obvious connection to survival. Why are couples possessive of each other? Strong emotions lead to death, killing, and celibacy. On the other hand, you're implying some kind of "superego" that can create emotions for a specific purpose (in early humans? in our genetic ancestors?). Is that what you're saying?

  On 2/25/2010 at 10:08 PM, theSun said:

of course i'm not suggesting any sort of predestination or "end point" of evolution, dunno how you got that. we're not "supposed" to do anything.

When you said "we didn't evolve to live in this technological world", it's easy to take that to mean "we weren't supposed to live with technology". This implies that there are ways we are supposed to live and not live. Encey is saying that the way we live now is perfectly natural because it happened; I agree.

  On 2/25/2010 at 10:08 PM, theSun said:

what does "justice" mean to a computer (getting back to the OP)? too complex, too abstract to define. also people of different cultures with different ideas of the self are bound to give completely unpredictable answers when asked what "justice" means. we don't all have the same perception of "justice" because it is so new to us, and a combination of common emotions based on circumstance. but something like fear, happiness or excitement we can all relate to, it's in our genes and is a byproduct of our evolution.

What makes you so sure we can all describe fear, love, or excitement in the same way? I think we can all relate to them, but if we were asked to define them we'd have more disagreements. To me this does not support your idea that emotions were created by us and are imprinted in our genetic material.

  On 2/25/2010 at 10:08 PM, theSun said:

also, every personal and social phenomena depends explicitly on the biological and chemical reactions required by the people experiencing said phenomena.

If you're saying that everyone's perception of everything is unique and unfathomable, I agree. Otherwise I have to ask you to restate this somehow.

  On 2/25/2010 at 10:08 PM, theSun said:

hope this clears some things up, i tend to be a little ambiguous because i don't often write about this shit.

You should totally write about it more! Your ideas are interesting. I would like to understand more but until you fill in some holes for me I'm having trouble agreeing with you on almost anything.

  On 2/26/2010 at 12:06 AM, A/D said:

Did you and Gordo read the same textbook or something? This seems to be mostly non sequitors and broad statements about some controversial and complex stuff, but you guys understand it. Maybe there's further reading you can suggest? Comments below if you care to elaborate.

 

  On 2/25/2010 at 10:08 PM, theSun said:

fear, love, hate were all developed unwittingly by our ancestors to give future generations a better chance at survival.

I don't get this. You're saying emotions were consciously created in order to give an evolutionary advantage? On the one hand, emotions do not share an obvious connection to survival. Why are couples possessive of each other? Strong emotions lead to death, killing, and celibacy. On the other hand, you're implying some kind of "superego" that can create emotions for a specific purpose (in early humans? in our genetic ancestors?). Is that what you're saying?

emotions weren't consciously created just like opposable thumbs weren't consciously created. they ABSOLUTELY have a connection to survival. fear can deter, happiness/lust can motivate, extreme panic can induce extreme feats of strength. if we didn't instinctually fear bears attacking us we'd be stupidly brave. there are also some tests (sorry i don't have an article) that suggest laughing is instinctual for humans.

  Quote

  On 2/25/2010 at 10:08 PM, theSun said:

what does "justice" mean to a computer (getting back to the OP)? too complex, too abstract to define. also people of different cultures with different ideas of the self are bound to give completely unpredictable answers when asked what "justice" means. we don't all have the same perception of "justice" because it is so new to us, and a combination of common emotions based on circumstance. but something like fear, happiness or excitement we can all relate to, it's in our genes and is a byproduct of our evolution.

 

What makes you so sure we can all describe fear, love, or excitement in the same way? I think we can all relate to them, but if we were asked to define them we'd have more disagreements. To me this does not support your idea that emotions were created by us and are imprinted in our genetic material.

fear is fairly close to universal i'd say. so is happiness. some people fear goats, they experience anxiety that can be physically defined (increase in certain hormones, certain areas of the brain firing, sweat). i don't fear goats, this doesn't mean i'm immune to fear just that we aren't all scared of the same things. in this way i believe it is universal, when an emotion can be physically defined biologically. something like "justice" is a human construct, too complex to analyze physically (at this point).