4 ms·
There are plenty of fixed assumptions within deep learning. Off the top of my head: (1) Loss function (2) Pooling layers (which hard code invariances)
by tagh 7y ago
There are plenty of fixed assumptions within deep learning. Off the top of my head:
(1) Loss function
(2) Pooling layers (which hard code invariances)