Does anyone can give me any document where a proper way of coarse-graining is presented for entangled polymer dynamics from atomistic simulations, without any input from the tube picture? In general, the projection operator idea realizes the coarse-graining, but we need proper mobility matrix that represents the entanglement constraint. I think it is not automatically obtained, specifically under large deformation / fast flows. The difficulty seems essentially common for the GENERIC formalism developed by ETH group...??