Re:Google AI invents its own cryptographic algorithm
by verbify on Friday October 28, 2016 @06:45PM
(#12821403)
Would it be possible/easy to add the speed of encrypting/decrypting the data as a separate loss function? Potentially this could lead to cryptography being a less expensive computation.
[ Parent
| on HN ]
Re:Google AI invents its own cryptographic algorithm
by dgacmu on Friday October 28, 2016 @09:27PM
(#12822118)
It could, but within a given neural network structure, the speed is going to pretty much be constant. (Barring optimizations such as eliminating zero weights/activations). There's a meta-level above this of trying to search or automatically determine a "good" NN structure that can accomplish the encryption & decryption. That too (determining an optimal NN structure for a problem) is a fascinating research question in its own right! :) In fact, it's one that Jeff Dean called out a while ago as one of the leading-edge questions for DNNs, IIRC.