v1: Structure-preserving emulators for spectral/strength-function LRT, implemented for nuclear QRPA.
(ex. Yukiya's GT and T1/2 data. Good place to test parameter generalization)
v2: Generalize to other methods that produce compatible spectral responses.
(ex. Aaron's He2 oscillator spectrum)
v3 (rough sketch): New response representations: LIT transforms, Lanczos continued fractions, reduced linear solves, time-domain response.
(ex. Francesca's discrete response)
- The emulator may generalize if both datasets can be represented as strength functions or folded spectral responses, albeit it may not perform as well as it's original structure. Effectively, does the reduced spectral representation learned for QRPA have enough flexibility to represent CC response functions?
v1: Structure-preserving emulators for spectral/strength-function LRT, implemented for nuclear QRPA.
(ex. Yukiya's GT and T1/2 data. Good place to test parameter generalization)
v2: Generalize to other methods that produce compatible spectral responses.
(ex. Aaron's He2 oscillator spectrum)
v3 (rough sketch): New response representations: LIT transforms, Lanczos continued fractions, reduced linear solves, time-domain response.
(ex. Francesca's discrete response)