Quantum Monte Carlo

Quantum Monte Carlo (QMC) methods aim to solve the many-body Schrödinger equation, a fundamental problem in physics and chemistry, by stochastically sampling the wavefunction of a quantum system. Current research heavily utilizes neural networks, such as FermiNet and Psiformer, as flexible variational ansätze within QMC frameworks, exploring novel optimization techniques like Wasserstein gradient flows to improve convergence and accuracy in calculating both ground and excited states. These advancements are enabling highly accurate simulations of molecular systems, including challenging calculations of excited states and properties like transition dipole moments, with potential applications in materials science and drug discovery.