Calvin Lab Auditorium
We consider the usual problem of circuit obfuscation: given some classical circuit C, can we prepare an obfuscated version which allows a user to simulate black-box access to C without learning anything else? Classically this problem is known to be impossible, even for relatively weak formulations. But it may still be possible to produce a quantum state which obfuscates C, even in the very strong sense that any measurement of that state can be simulated (up to computational indistinguishability) using black-box access to C.
I will describe a protocol for quantum obfuscation of classical circuits, whose security is open. Time permitting I will discuss some of the technical issues that come up in the analysis.