Idealized models are a useful tool for proving the security of cryptographic schemes. For example, the random oracle model (ROM) gives an idealization of crytographic hash functions, whereas the GGM and AGM give idealizations of cryptographic groups.
I recently gave a talk to a reading group on the paper "To Label, or Not To Label (in Generic Groups)" by Mark Zhandry. You can find my notes here.
