2 ms·

I think the method of creation is sometimes relevant. Like if you can prove that you wrote a poem before someone else did, but they registered the same work coi

I think the method of creation is sometimes relevant. Like if you can prove that you wrote a poem before someone else did, but they registered the same work coincidentally I don’t think you will be treated the same as if it is shown that you copied. So it’s more than if the results are sufficiently novel to a blinded human... perhaps in the future this will be an area where you could be audited - forced to show your training sets or you will be assumed guilty???

buboard 7y ago
I think there have been many, many similar copyright cases in music. E.g. composers have performed their composition method to prove originality. It would be surprising if judges started claiming that dataset companies have excusivity to all NN-derived works , and would seriously hold back progress.