From what I understand, the folklore way to batch Ring-LWE style cipher texts is to use the Chinese remainder theorem. I am wondering if there are any different approaches/optimizations to this style of (ring-LWE) batching and where I would be able to read about them.
Thank you!
There are some other ways to amortize computation in RLWE-type FHE. Halevi and Shoup have a technique they call hoisting which can be useful. Li, Micciancio, Raykova, and myself have a technique that we called composable preprocessing. Both techniques are somewhat specific (e.g. less general than using the CRT to get SIMD-type operations). I don't personally know of any broader survey of non-CRT amortization techniques though.
