Abstract: Training of autoencoders using the back-propagation algorithm is challenging for non-differential channel models or in an experimental environment where gradients cannot be computed. In this ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results
Feedback