Ive not read the whole paper that you are referring to, so its hard for me to say anything with too much confidence. The paper seems two involve multiple meshes with assembly from functions on one mesh being transferred to the other, which I guess one could do with the non-matching mesh interpolation implemented by @Massimiliano_Leoni (see for instance https://github.com/FEniCS/dolfinx/blob/84b16751821d359fb2c5db74bcf2894b9c69db74/python/test/unit/fem/test_function.py#L180).