Re:ARC-AGI without pretraining
by mmazing on Wednesday March 05, 2025 @12:18AM
(#43263134)
> But, the neural networks are pretty small, so you could train a "hypernetwork" that predicts μ, Σ, θ for a given puzzle, and even predicts how to train these parameters.
Kurt Godel (or maybe Douglas Hofstadter, rather) would raise an eyebrow. :)
Kurt Godel (or maybe Douglas Hofstadter, rather) would raise an eyebrow. :)