Though this breakthrough has not been achieved using quantum computing, it’s actually amazing and interesting in terms of machine learning techniques applied to quantum chemistry problems.
Researchers struggle to determine the exact wave function when analyzing a normal chemical molecule system, which has its nuclear position fixed and electrons spinning. They often use a Slater-Jastrow Ansatz application of quantum Monte Carlo (QMC) methods.
A team at DeepMind have brought QMC to a higher level with the Fermionic Neural Network — or Fermi Net — a neural network with more flexibility and higher accuracy. Fermi Net takes the electron information of the molecules or chemical systems as inputs and outputs their estimated wave functions, which can then be used to determine the energy states of the input chemical systems.
In their experiments the team proved the superior accuracy of Fermi Net compared to other QMC alternatives. (medium.com)