Detection of Hardware Trojans Using Controlled Short-Term Aging

> physical phenomenon of short-term aging as a route to detecting hardware Trojans

Does anyone understand how this is supposed to work? As near as I can understand it, the idea is that an IC with a trojan horse will behave correctly for some period of time, after which it'll flip over to doing something surreptitious. (For example, the IC might generate random 256-bit cryptographic keys for a long period, but then switch over to always setting 200 bits of the 256 bit key to a known value to make it easy for an adversary to break.) The short-term aging tries to accelerate the period in which the IC is behaving correctly, and get it to switch over sooner to its trojan horse behavior, so that the test bed can detect it. Is that the idea?