3 ms·
We are unconsciously (pun intended) implementing how brains work both in dream and wake states. Can't wait until we add some kind of (lossless) memory to this m
by thegabriele 2y ago
We are unconsciously (pun intended) implementing how brains work both in dream and wake states.
Can't wait until we add some kind of (lossless) memory to this models.
- hackernewds 2y agoAny evidence to back this lofty claim?
- sweeter 2y agovibes
- soraki_soladead 2y agoWe have lossless memory for models today. That's the training data. You could consider this the offline version of a replay buffer which is also typically lossless. The online, continuous and lossy version of this problem is more like how our memory works and still largely unsolved.