Algorithm produces one of the best solutions to molecules’ Schrödinger equations yet
But aggressive ‘pre-training’ actually damages its predictive powers
A new deep-learning algorithm from researchers in Austria produces more accurate numerical solutions to the Schrödinger equation than ever before for a number of different molecules at relatively modest computational cost. Surprisingly, the researchers found that, whereas some ‘pre-training’ of the algorithm could improve its predictive abilities, more substantial training was actively harmful.
As the Schrödinger equation can be solved analytically only for the hydrogen atom, researchers wishing to estimate energies of molecules are forced to rely on numerical methods. Simpler approximations such as density functional theory and the Hartree-Fock method, which is almost as old as the Schrödinger equation itself, can treat far-larger systems but often gives inaccurate results. Newer techniques such as complete active space self-consistent field (CASSCF) give results closer to experiments, but require much more computation.