Computers Are Incapable Of Creatively Writing Music

[cluelusshusbund]

A computer doesn't have a life and all the quirky emotions, feelings and addictions arising from encounters with this and that, as well mental illness and depression that drive the imagination of some artists.

Yes… humans... like AI... appear to have causes for ther actions.!!!

 

[C C]

Yes… humans... like AI... appear to have causes for ther actions.!!!

Why would they not? Remarking about that is like walking by a house with a group of people, pointing at the lawn, and shouting while jumping excitedly up and down: "OMG, the grass is green! It's green! Do you see that, mind-boggling! Somebody also mowed the grass -- what a trip, it's green and mowed!"

We don't expect a dog to do everything a human can do because the latter's body lacks certain items to yield the necessary chains of causation, and millennia of cultural developments, to conceive and output rocket ships. Similarly, current computers with a tabletop existence are likewise deficient in their own way, as well as missing an embodied life in an environment.

 

[C C]

Duplicate post deleted.

 

[Write4U]

"Emotion, feelings" as well as idiosyncratic habits, preferences, and interests developed as a result of individual life incidents -- as addressed on this segment of the thread -- does not refer to what a passive viewer or listener is personally experiencing. But how such may actively affect the choices/decisions of an artist, composer, novelist, etc in the course of their creating a work. Any idiot can master rote and work entirely within a strict, predictable framework or template. The surprises come from rebellious impulses and attractions that wander from that.

yes, but that does not address the composition itself, it addresses the performance, and I admit as the two musicians observed, the performance lacked the subtle nuances which is the role of the conductor to introduce that abstraction during the performance. And most conductors have a favorite composer and an affinity to their work.

Arturo Toscanini (1867-1957)
Many conductors today still consider Toscanini the greatest conductor of Verdi who ever lived. Not surprising, really, seeing as he actually began his musical life playing under Verdi himself. Toscanini also gave the première performances of 'La bohème' and 'Turandot', so he's basically a legend.

https://www.classicfm.com/discover-music/latest/great-conductors/arturo-toscanini/

Schumann

For the past 45 years I have been living, breathing and thoroughly absorbing the Schumann Symphonies. Having a special affinity with all four, I have conducted them with some 30 different orchestras around the world. During all these years, I have studied and researched everything I could find about Schumann. There is no doubt that the character of a musical work is determined by the basic choice of tempo. The same music performed in a flowing basic tempo has a totally different effect if performed in a halting slower one. Why then do so many conductors (as well as pianists and other performers) simply ignore the specific wishes and instructions of Schumann in that regard? We are well aware that Schumann was most particular in prescribing meticulous metronome markings in most of his works.

Moreover, he went back a few times later in his life to recheck his metronome markings and occasionally amend them. We also know from letters of his friends that while his wife Clara was reading some of his works at the piano, Robert stood behind her, lightly tapping the correct tempo on her shoulder, to ensure that she didn’t play it too slowly! As Schumann was what one would call “a fast-pulse person” his agitated metronome markings clearly reflect his personality and character. As this article deals mainly with the composer’s four symphonies, I would like to focus on explaining the utmost importance of his metronome markings for the overall understanding of these great works. I am well aware that the validity of Schumann’s metronome markings (henceforth m.m.) had been subject to dispute as to how far they should be binding (as were Beethoven’s).

https://myscena.org/yoav-talmi/conductors-orchestras-metronome-markings-in-schumanns-symphonies/

i.e. it's the performance that makes the notes come alive. Make sheet music of the AI composition and let a live orchestra perform, that will complete the "unfinished" symphony.

 

[cluelusshusbund]

Why would they not?

Do you thank that the "personal inspirations" of a human painter are any less programed than the output of a programed AI painter.???

 

[Write4U]

Well for you music lovers. Test your musical acumen against these samples;
1 composed by human and 1 composed by AI.

 

[C C]

Do you thank that the "personal inspirations" of a human painter are any less programed than the output of a programed AI painter.???

Somebody wrote a program and installed it via one of the ports on Hieronymus Bosch's skull, rather than his predilections arising from genetics, rearing, and social and environmental interactions?

Even what today's AI is grounded in doesn't depend exclusively upon a human designing software for them: "Machine learning involves computers discovering how they can perform tasks without being explicitly programmed to do so. It involves computers learning from data provided so that they carry out certain tasks."

Still, there's a limit to what a tabletop proto-intelligence without senses and a body can apprehend, just as there's a limit for a human child confined to and immobilized in a room for years: https://en.wikipedia.org/wiki/Genie_(feral_child)

 

[C C]

- - - - - delete - - - - -

 

[Write4U]

Still, there's a limit to what a tabletop proto-intelligence without senses and a body can apprehend, just as there's a limit for a human child confined to and immobilized in a room for years: https://en.wikipedia.org/wiki/Genie_(feral_child)

I agree and so do the AI developers. The additional power of GPT3 lies in its vastly expanded RAM functions, and they have bigger plans yet.
AFAIK, the only limitation lies in capacity to handle terrabites of data

Which Transformer Should I Go With: GTP-2 or GPT-3?

The Generative Pre-Trained Transformer (GPT) is an innovation in the Natural Language Processing (NLP) space developed by OpenAI. These models are known to be the most advanced of its kind and can even be dangerous in the wrong hands. It is an unsupervised generative model which means that it takes an input such as a sentence and tries to generate an appropriate response, and the data used for its training is not labelled.

GPT-2 is an unsupervised deep learning transformer-based language model created by OpenAI back in February 2019 for the single purpose of predicting the next word(s) in a sentence. GPT-2 is an acronym for “Generative Pretrained Transformer 2”. The model is open source, and is trained on over 1.5 billion parameters in order to generate the next sequence of text for a given sentence. Thanks to the diversity of the dataset used in the training process, we can obtain adequate text generation for text from a variety of domains. GPT-2 is 10x the parameters and 10x the data of its predecessor GPT.

Simply put, GPT-3 is the “Generative Pre-Trained Transformer” that is the 3rd version release and the upgraded version of GPT-2. Version 3 takes the GPT model to a whole new level as it’s trained on a whopping 175 billion parameters (which is over 10x the size of its predecessor, GPT-2). GPT-3 was trained on an open source dataset called “Common Crawl”, and other texts from OpenAI such as Wikipedia entries.

IOW, basically unlimited access to any open source on the internet

GPT-3 was created to be more robust than GPT-2 in that it is capable of handling more niche topics. GPT-2 was known to have poor performance when given tasks in specialized areas such as music and storytelling. GPT-3 can now go further with tasks such as answering questions, writing essays, text summarization, language translation, and generating computer code. The ability for it to be able to generate computer code is already a major feat unto itself. You can view some GPT-3 examples here.

For a long time, many programmers have been worried at the thought of being replaced with artificial intelligence and now that looks to be turning into reality. As deepfake videos gain traction, so too is speech and text driven by AI to mimic people. Soon it may be difficult to determine if you’re talking to a real person or an AI when speaking on the phone or communicating on the Internet (for example, chat applications).

.....more
https://www.exxactcorp.com/blog/Deep-Learning/gpt2-vs-gpt3-the-openai-showdown

From the above examples I believe the GPT2 limitation in music composing has been pretty well solved. I will concede that it takes human performance to evoke maximum emotional response, leading to thoughts of the noblest kind.

Human brain

A typical computer runs on about 100 watts of power. A human brain, on the other hand, requires roughly 10 watts. That’s right, your brain is ten times more energy-efficient than a computer. The brain requires less power than a lightbulb.

Conclusion

Ultimately, there is no clear winner overall. Human beings and computers have their own advantages, depending on the category. If you want precision and raw processing speed, a computer is the clear choice. If you want creativity, energy efficiency, and prioritization, a human is your best bet.

The good news is that we don’t have to choose. It doesn’t have to be a contest of humans against computers. We can work together and enjoy the best of both worlds. That is, until Skynet becomes self-aware.

https://www.crucial.com/blog/technology/how-does-the-human-brain-compare-to-a-computer#

 

[cluelusshusbund]

Do you thank that the "personal inspirations" of a human painter are any less programed than the output of a programed AI painter.???

Somebody wrote a program and installed it via one of the ports on Hieronymus Bosch's skull, rather than his predilections arising from genetics, rearing, and social and environmental interactions?

Do you thank that the “Sombody” who wrote the program had also been programed to write the program that they wrote.???

Even what today's AI is grounded in doesn't depend exclusively upon a human designing software for them: "Machine learning involves computers discovering how they can perform tasks without being explicitly programmed to do so. It involves computers learning from data provided so that they carry out certain tasks."

Still, there's a limit to what a tabletop proto-intelligence without senses and a body can apprehend...

Do you thank its not posible for AI to evolve to a pont of havin senses which will allow it to apprehend like HI.???

 

[C C]

Do you thank that the “Sombody” who wrote the program had also been programed to write the program that they wrote.???

Seems similar to the homunculus fallacy or the Droste effect, of reiteration that goes on ad nauseam. At some point you have to introduce a different provenance than a repeat of _X_, to end the nested Russian dolls scenario.

Do you thank its not posible for AI to evolve to a pont of havin senses which will allow it to apprehend like HI.???

If we don't exterminate ourselves or go Neo-Luddism before then, I expect self-replicating machines to replace us eventually and slowly migrate across the galaxy like technological wildlife. Regardless of whether either human engineers or machine learning innovations ever figure out how to do endow them with everything we might want to. (For instance, aspects of complex life like "progressive sapience" and "extrospective and introspective experiences" can't match the survival tenacity of dumb, p-zombie microbes. Certain traits are not necessary to succeed in a proliferation sense.) But I wouldn't be surprised if an archailect emerges at some point over the centuries (though not as ludicrously swift as Raymond Kurzweil expects).

If there could be such a thing as a non-cognitive based "purpose" for something like humans, that would seem to qualify as it: To implement the next step that can tolerate and thrive in the greater hostile environment of the universe. We're a necessary mediator for that, since the traditional stratum of biological evolution has never successfully adapted to space. If microscopic life has traveled to other planetary environments in the solar system via panspermia, it seems to be largely dormant or not indulging in anything impressive. (And lots of luck with any multicellular marine stuff from Europa or Enceladus breaking through the frozen crust as macroscopic space nomads.)

 

[cluelusshusbund]

Do you thank that the “Sombody” who wrote the program had also been programed to write the program that they wrote.???

Seems similar to the homunculus fallacy or the Droste effect, of reiteration that goes on ad nauseam. At some point you have to introduce a different provenance than a repeat of _X_, to end the nested Russian dolls scenario.

“ad nauseam” not necessary... the period of time after the big-bang will cover it… so wit that cleared up i will guess that you’r answr to my question is yes.!!!

Do you thank its not posible for AI to evolve to a pont of havin senses which will allow it to apprehend like HI.???

If we don't exterminate ourselves or go Neo-Luddism before then, I expect self-replicating machines to replace us eventually and slowly migrate across the galaxy like technological wildlife. Regardless of whether either human engineers or machine learning innovations ever figure out how to do endow them with everything we might want to.

Thanks... i will take that as a yes in answer to my queston.!!!

 

[cluelusshusbund]

Edit to above post in red.!!!

↑
cluelusshusband --- Do you thank its not posible for AI to evolve to a pont of havin senses which will allow it to apprehend like HI.???
↑
CC --- If we don't exterminate ourselves or go Neo-Luddism before then, I expect self-replicating machines to replace us eventually and slowly migrate across the galaxy like technological wildlife. Regardless of whether either human engineers or machine learning innovations ever figure out how to do endow them with everything we might want to.

cluelusshusbund --- Thanks... i will take that as a no in answer to my queston.!!!

 

[Steve Klinko]

Question is if you would hang it on your wall as an original? I would. An AI painting from the Impressionist school.

Sure, but that doesn't mean there was any Creativity in it's production.

 

[Steve Klinko]

Nope. Programs intended for art do not use "computer algorithmic calculation." They learn what good art is and create new art based on that learning.

Computers can only incorporate Pattern Matching into their Algorithms. The Pattern Matching is implemented by Neural Nets. Computers don't Learn how to do anything but they can be Configured to do things.

 

[Vociferous]

Omitting the copious straw men that I've already repeatedly called out...

Now all you have to do is show how any goal is not merely a subjective matter (which relegates creativity not to the mechanics at work but to a matter of perception). Until you can, it is your own argument that results in nothing being creative.

You're crackpot "prove me wrong' is sad and boring. You're the one making the claim that AI can be creative. So any argument you make about creativity being merely subjective or nonexistent is just you refuting your own claim. If that's really what you want to do, I'm fine with that.

Yes, there can: if your argument is that something complex (humans) can create, but never explain how that happens, then you are simply appealing to that complexity. Period.

Again, your claim, so your onus to show how what you claim of AI is comparable to that in humans. If you think there's a problem with the relative complexity of the two, it's on you to solve that.

No, the null hypothesis is that there is no difference between certain characteristics of a population. So if you make an argument based on behaviour within a ruleset, the null hypothesis is that anything else in a ruleset will behave the in a same way with respect to that ruleset. E.g. if one system follows the ruleset it is governed by, another system will follow the ruleset that it is governed by. If you want to reject the null hypothesis you need to make a case for it, a case that is conspicuous by its ongoing absence.

Oh, it's so comically sad. You literally say that "the null hypothesis is that there is no difference between certain characteristics of a population", but then immediately claim there's no difference between different populations, explicitly ignoring any of their specific characteristics.

As already clarified, that was in comparing us to AI, not in any absolute sense.

Yes, you seem very intent on avoiding any such comparison, as it's unfavorable to your claim.

You really don't understand how something like the first airplane or a work of art is new?

I'm not doing your homework for you. You've claimed they're new, so support the claim. Your arguments to-date would suggest that they are not new, that they are just innovation on existing ideas, which you have consigned to the camp of "not creative". So show how they are new, or is all you have now the common refrain of the crackpot: "Prove me wrong!" Show that they are new. Don't just assert it. Examine how they arose with your own argument in mind. Show how, following your argument, they should be classified as new, given what you have rejected thus far and why. Unless you can do that you simply have nothing. And remember, these are your arguments being interrogated here.

No homework. If you really can't understand how those things are new, again, you're only defeating your own claim. I'm happy to let you.

I'm waiting for you to provide the difference beyond an appeal to complexity. Can you do that? Or are you just going to continue to evade?

Do you really think this sort of crackpot trolling is making any point?

No, it's not. It's the conclusion of your arguments once you strip out the appeals to complexity. If you can't or won't show why things don't apply to one system but do apply to a more complex one, you are appealing to complexity. I get that you want your arguments to show one can be creative and the other not, but your arguments sans your appeal to complexity simply don't do that.

You're the one making the positive claim (the only kind that can be evidenced), that AI can be creative. Apparently you can't manage to distinguish creativity from innovation without refuting creativity, so you just project that on others. Yawn.

We can get on to my arguments in this regard once we have finished with yours. And who knows, by examining yours further you may yet begin o understand mine. But for now we're examining what you have to say, and have said, about the matter.

Another refrain of the wild crackpot in its natural habitat. We can only be blessed to hear the argument for your positive claim after we "prove you wrong" by proving a negative, which can't be done. The null hypothesis that there are no pink unicorns can never be exhaustively proven. But if you think you've found one, in the form of AI creativity, that's the only claim that can be demonstrated. So all this crackpottery is just endless evasion and straw men.

I'm sure we'll get on to that in due course.

Won't be holding my breath.

When I claim something the onus will be on me, and I will own it.

Wait, so you've never claimed that AI can be creative? Okay, we're done here.

Cheers.

 

[Sarkus]

Omitting the copious straw men that I've already repeatedly called out...

Your continued evasion is noted, this time through simply ascribing "strawman" to anything you don't want to answer. Typical bullshit one comes to expect from you.

You're crackpot "prove me wrong' is sad and boring.

Ironic strawman from you, given that I am not saying "prove me wrong" but rather asking you to prove your position. Note the difference?

You're the one making the claim that AI can be creative.

I have made that claim, but my current "discussion" with you is separate to that, as it looks at your arguments, your claims, and shows how they lead to the conclusion that creativity doesn't exist. Deal with that or don't, but stop being a dick and trying to turn the onus onto me, when it's clearly you who is failing to argue the conclusion you want to reach.

So any argument you make about creativity being merely subjective or nonexistent is just you refuting your own claim. If that's really what you want to do, I'm fine with that.

No, it's all part of your argument: you would need to show that goals are not a matter of subjectivity, etc. If you can't accept that then you're simply being dishonest.

Again, your claim, so your onus to show how what you claim of AI is comparable to that in humans.

So you're happy with your appeal to complexity, is that it? You're not able/willing to support your position in that regard? If so, just say so, it would be a lot more honest that your current pathetic nonsense.

If you think there's a problem with the relative complexity of the two, it's on you to solve that.

I don't think there's a problem: one is more complex than the other. So what? It is now up to you to show that this difference in complexity gives rise to creativity etc. At the moment you are just claiming it. You're just appealing to that complexity. And you think that is honest?

Oh, it's so comically sad. You literally say that "the null hypothesis is that there is no difference between certain characteristics of a population", but then immediately claim there's no difference between different populations, explicitly ignoring any of their specific characteristics.

The population in this regard is anything that is governed by a ruleset. In the issue of rulesets, if A behaves according to its ruleset, then the null hypothesis would be that B behaves according to its ruleset. Or are you simply too thick to realise that? Yeah, it's comically sad: your efforts at rebuttal and avoidance.

Yes, you seem very intent on avoiding any such comparison, as it's unfavorable to your claim.

If you want to take the discussion down that path, feel free, but I was simply clarifying to you the distinction I was making. But hey, let's be dishonest and turn that into an unwillingness to avoid comparisons that were never made, shall we.

No homework. If you really can't understand how those things are new, again, you're only defeating your own claim. I'm happy to let you.

Then you really are being dishonest. Thanks for clarifying that you have no intention here other than to state your claim and then ignore all points made against it.
You highlighted the example as being of something new, and you are now failing to show that it is something new even under your own arguments. And you expect people to take you seriously?

Do you really think this sort of crackpot trolling is making any point?

It's not crackpot trolling to call out shit arguments for being shit, and the person who raised them as being devoid of any honesty. And that's all I'm doing here. Your arguments. Your inability to show how they lead to the conclusion you want, beyond your appeals to complexity, and your constant evasion.

You're the one making the positive claim (the only kind that can be evidenced), that AI can be creative.

Again, more dishonesty from you, trying to turn the current "discussion" we're having away from your shit arguments (that don't conclude as you think they do) and instead to other people's claims. You have claimed that AI can't be creative. You have yet to support it beyond arguments that lead to the absence of creativity entirely. You have appealed to complexity. You have raised strawmen. If you want to discuss my claim, that's fine, we can do that in due course. But for now, be honest. That's all I ask. Can you do that?

Apparently you can't manage to distinguish creativity from innovation without refuting creativity, so you just project that on others.

I'm just following your arguments, your definitions etc. I mean, if you want to simply assert the difference while your arguments themselves can't distinguish, and if you want to continue fallacious arguments, that's fine 'n' all.

Another refrain of the wild crackpot in its natural habitat.

Says the one looking to evade from his debacle of an argument. I'm just trying to keep you honest, but you are certainly strong in the dishonesty-force.

We can only be blessed to hear the argument for your positive claim after we "prove you wrong" by proving a negative, which can't be done.

You continue this strawman as if you're wedded to it. For the last time, I'm not asking you to prove a negative, only to support the claims you make, and show that the arguments you make actually lead to the conclusions you think they do, without, that is, appealing to complexity etc. Can you do that? These are your arguments you're making, not mine. Support them. Please. Or is it a case of "I'm claiming a negative therefore I don't have to even show that my arguments lead to the conclusions I think they do!"?

The null hypothesis that there are no pink unicorns can never be exhaustively proven.

But you can at least set out an argument that leads to the conclusion that we shouldn't take their existence as anything other than akin to imagination. Unfortunately that's where your analogy breaks down, as your argument for the lack of creativity in AI leads to the conclusion, as shown, that creativity doesn't exist. You accuse me of arguing it away entirely but they are your arguments. I'm quite happy to agree that if there is no creativity then AI are not creative. That's all you've shown with your arguments, thus far.

But if you think you've found one, in the form of AI creativity, that's the only claim that can be demonstrated. So all this crackpottery is just endless evasion and straw men. Won't be holding my breath.

Calling someone else's dismantling of your argument "crackpottery" certainly gives me a laugh. It confirms your tendency to evade, and highlights your dishonesty. If you want to have an actual discussion, let me know, but you're going to have to do better.

Wait, so you've never claimed that AI can be creative? Okay, we're done here.

Oh, I have, but that's not the discussion we've been having. We've been discussing your arguments as to why AI aren't creative. If you can't be honest enough to realise that... ah, well.

 

[billvon]

Computers can only incorporate Pattern Matching into their Algorithms.

Nope. Neural networks can do far more than pattern matching.

Computers don't Learn how to do anything but they can be Configured to do things.

This book may help you with your confusion:

Neural Networks and Deep Learning
http://neuralnetworksanddeeplearning.com/

 

[Steve Klinko]

Nope. Neural networks can do far more than pattern matching.

This book may help you with your confusion:

Neural Networks and Deep Learning
http://neuralnetworksanddeeplearning.com/

Thank You. I know what Deep Learning is. It is just a Neural Net with lots more layers made possible by faster Computers and more efficient Algorithms. I hope you don't think that just because the Book says Deep Learning, that Computers are really Learning. They are electronic devices that can be Configured. We only say they Learn by convention and convenience. There is no Learning in the same sense as what a Human Brain does.

 

[Write4U]

Sure, but that doesn't mean there was any Creativity in it's production.

Does that matter? How do you define creativity? How many truly creative artists are there? Picasso, Daly, Bosch ?

I think that example of a "windy day" showing the leaves drifting in the air (all in the same direction) is very creative. I cannot recall ever seeing anything like it.