Abstract: In this informal talk I will describe the Quantum Monte Carlo (QMC) methods implemented in QMCPACK, the scientific motivations for developing these methods further, and highlight some of the key remaining challenges. Besides possible improvements in the fundamental algorithms and methodology, the statistical nature of the method results in several unsolved problems where robust optimization in the presence of noise is required, as well as in postprocessing (denoising) of the computed properties. Much cheaper, but less accurate, deterministic methods can be applied to all to systems where QMC is of interest, potentially acting as a source of reference data. The software is currently supported as part of the Exascale Computing Project as well as a BES funded Computational Materials Sciences Center. The goal of this talk is to identify possible collaborations and problems of join interest.
Bio: Paul Kent is distinguished research staff in CSED and CNMS. He is a Fellow of the American Physical Society and author of over 150 publications with 35000 citations.
Last Updated: May 28, 2020 - 4:01 pm