@haritulsidas Neither the linked article, nor the paper mention anything about ML or an 80% prediction. Where did you get those numbers? This only has real implications for cryptography if cryptographic systems rely on physical coin tosses to generate random numbers.
@solalnathan Or a fancy way to perform data augmentation?
@metin@mastodon.art I believe that you should be able, at least in theory, to generate your artwork based on its description using the final trained model. By description I mean either the text description you add while uploading the work or a an actual natural language description of the piece (or a combination of the two). If you get your artwork as output, that might be a good signal. But that depends entirely on the UX they will build around the model and it is not scaleable
Exploring ways of applying machine learning to address climate change.
Background in computer science and machine learning + team management and leadership.