This ML paper seems potentially relevant; it trains a neural network to find features where you can improve your predictions of information at a time distance by assuming the features to be better conserved.
Nice! That is a pretty good fit for the sorts of things the Telephone Theorem predicts, and potentially relevant information for selection theorems as well.
This ML paper seems potentially relevant; it trains a neural network to find features where you can improve your predictions of information at a time distance by assuming the features to be better conserved.
Nice! That is a pretty good fit for the sorts of things the Telephone Theorem predicts, and potentially relevant information for selection theorems as well.