3 ms·
I read the whole paper, but I still don't get exactly why this relation exists. They described some properties of "memory systems", and then said that a memory
by wyager 13y ago
I read the whole paper, but I still don't get exactly why this relation exists. They described some properties of "memory systems", and then said that a memory system must "remember more than one thing", and for some reason knowing the future as well as the past corresponds to "only knowing one thing", which means that anything that knows the future isn't a memory, because it only knows one thing.
Can someone please explain the whole thing with "knowing multiple things" and robustness to perturbations in the system?