7 ms·

I find the idea that weights are not copyrightable very fascinating - appealing even. I have a hard time imagining a world where this is the case, though. Can

I find the idea that weights are not copyrightable very fascinating - appealing even. I have a hard time imagining a world where this is the case, though. Can you summarize why weights would not be copyrightable or give me pointers to sources that support that view.

slimsag 3y ago
Speculating (I am not a lawyer) I see two options: 1. Model weights are the output of mathematical principles, in the US facts are not copyrightable, so in general math is not copyrightable. 2. Model weights are the derivative work of all copyrighted works it was trained on - in which case, it would be similar to creating a new picture which contains every other picture in the world inside of it. Who is the copyright owner? Well, everyone, since it includes so many other copyright holders' works in it.
- humanistbot 3y ago
  Your second argument, if true, disproves your first argument.
  - slimsag 3y ago
    Doesn't matter. A court decides in the end, and the two choices I presented could lead to OPs scenario. If a court decides that, they decide that, period. I'm not 'making an argument' with those points - I'm presenting options a court might choose from when setting precedent.
- earleybird 3y ago
  Your second question asks: "Who owns the Infinite Library[0]?" related, there was a presentation (i've lost the reference) on automatic song (tune?) generation where the presenter claimed (rather humourusly) that he'd generated all the songs that had ever been and will ever be so that while he was infringing on a large but finite number of songs, he was non infringing on an infinite number of future songs. So, on balance he was in a favourable position. [0] https://en.wikipedia.org/wiki/The_Library_of_Babel https://en.wikipedia.org/wiki/The_Library_of_Babel
- FishInTheWater 3y ago
  Remember that database rights are a thing. One cannot hold copyright facts, but one can "copyright" a collection of facts like a search index or a map.
sebzim4500 3y ago
Generally the output of a machine is not copyrightable. Similarly, the contents of a phone book is not copyrightable in the US even if the formatting/layout is. So I could take a phonebook and publish another one with identical phone numbers as long as I laid it out slightly differently.
- cal85 3y ago
  What about compiled binaries? If I write my own original source code (and thus automatically own the copyright to it), and compile it to binary, is the binary not protected to?
  - sebzim4500 3y ago
    No, because you the input to that process was a bunch of work that you did. In the case of an LLM, I don't think that the work of compiling the training data probably would qualify by analogy to the phonebook example.
    - cal85 3y ago
      Sure, but I was just responding to "Generally the output of a machine is not copyrightable", which seemed obviously wrong to me... But on reflection, you are totally right, I was just getting mixed up on the distinction between copies and the creative works themselves. Machine output of something is generally just a copy of something. Whatever it is a copy of may be a copyrightable work, and if so, whoever came up with that original work has the right to all the copies output by machines (or copies generated by hand-tracing, or whatever). Anyway, on LLMs... Even if we assume LLM weights are just copies (machine outputs) of whatever inputs they were trained on, then I assume I would automatically own the exclusive right to restrict the distribution of weights of a 'Me' chatbot trained exclusively on my own writings. But what if someone else comes along and writes a load of bespoke code specifically to generate improved weights for this same model, so the resultant chatbot works much better in conversation (still with my tone of voice, but with better performance and better interpretation of questions)? Is that programmer not adding some creative value, such that we might both have a right to restrict distribution of those improved weights? (NB. it's common for an item to be a 'copy' of multiple original works, e.g. copies of Jimi Hendrix's cover of Bob Dylan's 'All Along the Watchtower'.)
paxys 3y ago
It's a complicated question and I don't think anyone can give a clear yes or no answer before some court has ruled on it. One school of thought is that copyright is designed to protect original works of creativity, but weights are generated by an algorithm and not direct human expression. AI generated art, for example, has already been ruled ineligible for copyright.
appplication 3y ago
Let’s take a simple linear regression model with a handful of parameters. The weights could be an array of maybe 5 numbers. Should that be copyrightable? What if someone else uses the same data sources (e.g. OSS data sets) and architecture and arrives at the same weights? Is this a Copyright violation? Let’s talk about more complex models. What if my model shares 5% of the same weights with your model? What about 50%? What about 99%? How much do these have to change before you’re in the clear? What if I take your exact model and run it through some extra layers that don’t do anything, but dilute the significance of your weights? It’s a murky area, and I’m inclined to think copyright is not at all the right tool to handle the legality of these models (especially given the glaring irony they are almost all trained using copyrighted material). Patents, perhaps better suited, but I’m also not sold.
cbm-vic-20 3y ago
An analog to this might be the settings of knobs and switches for an audio synthesizer, or guitar effects settings. If you wanted to get the "Led Zeppelin sound" from a guitar, you could take a picture of the knobs on the various pedals and their configuration, and replicate that yourself. You then create a new song that uses those settings. Is that something that is allowed under copyright? What if there were billions of knobs, tuned after years of feedback and observations of the sound output?
- paxys 3y ago
  I don't think that's a good analogy. A piano has N keys. You can press certain ones in certain combinations and write it down. That result is still copyrightable, because you can prove that it was an original and creative work. Setting knobs for a machine is no different, but the key differentiator is if you did it yourself or if an algorithm did it for you.
  - cbm-vic-20 3y ago
    In my analogy, it's not the sequence of the notes or the composition, which I agree is copyrightable. But are the settings of the knobs and switches on synthesizers and effects devices used in a recording equivalent to the weights of a neural network or LLM? And if so, are those settings or weights copyrighitable?
- rvcdbn 3y ago
  That’s a bad analogy because a human chose the values of those settings using their creative mind. That’s not at all the case with weights. This originality is the heart of copyright law.
rvcdbn 3y ago
I have a hard time imagining a world where it is not the case at least in the US i.e. where copyright is extended to a work with no originality in direct contradiction to copyright clause in the constitution.
- bilbo0s 3y ago
  It's all kind of irrelevant. If they are not copyrightable, then most companies will simply hide them behind an API. There is no law saying these companies must release their weights. The companies are releasing their weights because they felt they could charge for and control other things. Like the output from their models. If they can't charge for and control those other things, then we'll likely see far fewer companies releasing weights. Most of this stuff will move behind APIs in that scenario.
  - rvcdbn 3y ago
    Maybe, maybe not. Companies are not monoliths. For all we know, internally it’s already well known that model weights likely aren’t copyrightable and the only reason for the restrictions is to give the appearance of being responsible to appease the AI doomers.

slimsag 3y ago

Speculating (I am not a lawyer) I see two options: 1. Model weights are the output of mathematical principles, in the US facts are not copyrightable, so in general math is not copyrightable. 2. Model weights are the derivative work of all copyrighted works it was trained on - in which case, it would be similar to creating a new picture which contains every other picture in the world inside of it. Who is the copyright owner? Well, everyone, since it includes so many other copyright holders' works in it.

humanistbot 3y ago
Your second argument, if true, disproves your first argument.
- slimsag 3y ago
  Doesn't matter. A court decides in the end, and the two choices I presented could lead to OPs scenario. If a court decides that, they decide that, period. I'm not 'making an argument' with those points - I'm presenting options a court might choose from when setting precedent.
earleybird 3y ago
Your second question asks: "Who owns the Infinite Library[0]?" related, there was a presentation (i've lost the reference) on automatic song (tune?) generation where the presenter claimed (rather humourusly) that he'd generated all the songs that had ever been and will ever be so that while he was infringing on a large but finite number of songs, he was non infringing on an infinite number of future songs. So, on balance he was in a favourable position. [0] https://en.wikipedia.org/wiki/The_Library_of_Babel https://en.wikipedia.org/wiki/The_Library_of_Babel
FishInTheWater 3y ago
Remember that database rights are a thing. One cannot hold copyright facts, but one can "copyright" a collection of facts like a search index or a map.

sebzim4500 3y ago

Generally the output of a machine is not copyrightable. Similarly, the contents of a phone book is not copyrightable in the US even if the formatting/layout is. So I could take a phonebook and publish another one with identical phone numbers as long as I laid it out slightly differently.

cal85 3y ago
What about compiled binaries? If I write my own original source code (and thus automatically own the copyright to it), and compile it to binary, is the binary not protected to?
- sebzim4500 3y ago
  No, because you the input to that process was a bunch of work that you did. In the case of an LLM, I don't think that the work of compiling the training data probably would qualify by analogy to the phonebook example.
  - cal85 3y ago
    Sure, but I was just responding to "Generally the output of a machine is not copyrightable", which seemed obviously wrong to me... But on reflection, you are totally right, I was just getting mixed up on the distinction between copies and the creative works themselves. Machine output of something is generally just a copy of something. Whatever it is a copy of may be a copyrightable work, and if so, whoever came up with that original work has the right to all the copies output by machines (or copies generated by hand-tracing, or whatever). Anyway, on LLMs... Even if we assume LLM weights are just copies (machine outputs) of whatever inputs they were trained on, then I assume I would automatically own the exclusive right to restrict the distribution of weights of a 'Me' chatbot trained exclusively on my own writings. But what if someone else comes along and writes a load of bespoke code specifically to generate improved weights for this same model, so the resultant chatbot works much better in conversation (still with my tone of voice, but with better performance and better interpretation of questions)? Is that programmer not adding some creative value, such that we might both have a right to restrict distribution of those improved weights? (NB. it's common for an item to be a 'copy' of multiple original works, e.g. copies of Jimi Hendrix's cover of Bob Dylan's 'All Along the Watchtower'.)

paxys 3y ago

It's a complicated question and I don't think anyone can give a clear yes or no answer before some court has ruled on it. One school of thought is that copyright is designed to protect original works of creativity, but weights are generated by an algorithm and not direct human expression. AI generated art, for example, has already been ruled ineligible for copyright.

appplication 3y ago

Let’s take a simple linear regression model with a handful of parameters. The weights could be an array of maybe 5 numbers. Should that be copyrightable? What if someone else uses the same data sources (e.g. OSS data sets) and architecture and arrives at the same weights? Is this a Copyright violation? Let’s talk about more complex models. What if my model shares 5% of the same weights with your model? What about 50%? What about 99%? How much do these have to change before you’re in the clear? What if I take your exact model and run it through some extra layers that don’t do anything, but dilute the significance of your weights? It’s a murky area, and I’m inclined to think copyright is not at all the right tool to handle the legality of these models (especially given the glaring irony they are almost all trained using copyrighted material). Patents, perhaps better suited, but I’m also not sold.

cbm-vic-20 3y ago

An analog to this might be the settings of knobs and switches for an audio synthesizer, or guitar effects settings. If you wanted to get the "Led Zeppelin sound" from a guitar, you could take a picture of the knobs on the various pedals and their configuration, and replicate that yourself. You then create a new song that uses those settings. Is that something that is allowed under copyright? What if there were billions of knobs, tuned after years of feedback and observations of the sound output?

paxys 3y ago
I don't think that's a good analogy. A piano has N keys. You can press certain ones in certain combinations and write it down. That result is still copyrightable, because you can prove that it was an original and creative work. Setting knobs for a machine is no different, but the key differentiator is if you did it yourself or if an algorithm did it for you.
- cbm-vic-20 3y ago
  In my analogy, it's not the sequence of the notes or the composition, which I agree is copyrightable. But are the settings of the knobs and switches on synthesizers and effects devices used in a recording equivalent to the weights of a neural network or LLM? And if so, are those settings or weights copyrighitable?
rvcdbn 3y ago
That’s a bad analogy because a human chose the values of those settings using their creative mind. That’s not at all the case with weights. This originality is the heart of copyright law.

rvcdbn 3y ago

I have a hard time imagining a world where it is not the case at least in the US i.e. where copyright is extended to a work with no originality in direct contradiction to copyright clause in the constitution.

bilbo0s 3y ago
It's all kind of irrelevant. If they are not copyrightable, then most companies will simply hide them behind an API. There is no law saying these companies must release their weights. The companies are releasing their weights because they felt they could charge for and control other things. Like the output from their models. If they can't charge for and control those other things, then we'll likely see far fewer companies releasing weights. Most of this stuff will move behind APIs in that scenario.
- rvcdbn 3y ago
  Maybe, maybe not. Companies are not monoliths. For all we know, internally it’s already well known that model weights likely aren’t copyrightable and the only reason for the restrictions is to give the appearance of being responsible to appease the AI doomers.