[Cryptography] "Google teaches ???AIs??? to invent their own crypto and avoid eavesdropping"

[John-Mark Gurney]

Jerry Leichter wrote this message on Fri, Oct 28, 2016 at 18:02 -0400:
> Summary:  Alice, Bob, and Eve are neural nets; Alice and Bob share a key.  Alice tries to send messages to Bob.  Bob is rewarded for correctly reading the messages; Eve is likewise rewarded for correctly intercepting them; Alice is rewarded when Eve fails (and Bob succeeds, I guess).
> 
> In the experiments reported, Alice and Bob turn out to be pretty good at fooling Eve, who proves to be a mediocre cryptanalyst.

You know. designing an algorithm to best Eve when Eve shares w/ you
how well she is doing is pretty easy/nice.

I wish the NSA/[insert another attacker here] would be so kind.

-- 
  John-Mark Gurney				Voice: +1 415 225 5579

     "All that I will do, has been done, All that I have, has not."