ML-DFT Project at CASUS

--- id: 202008261735 --- # ML-DFT Project at CASUS ## Notes from last meeting ### Project 1 - DFT and Warm Dense Matter - Replace KS equations with ML models - Encode ionic structure (SNAP, ref below) - Feed forward neural nets, LDOS ### Project 2 - Physics Informed NNs - To accelerate KS equation - Specifically time dependent KS ## Literature Review ### DFT - The ABC of DFT - K. Burke [https://dft.uci.edu/doc/g1.pdf](https://dft.uci.edu/doc/g1.pdf) (check references too) - Sherrill Group Notes [http://vergil.chemistry.gatech.edu/notes/](http://vergil.chemistry.gatech.edu/notes/) - Density Functional Theory - D. Sholl, J. Steckel [https://www.onlinelibrary.wiley.com/doi/abs/10.1002/anie.200905551](https://www.onlinelibrary.wiley.com/doi/abs/10.1002/anie.200905551) - Time-dependent density functional theory - Marques M.A.L, Gross E.K.U - [link](https://www.annualreviews.org/doi/pdf/10.1146/annurev.physchem.55.091602.094449) ### ML - Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations - [https://www.sciencedirect.com/science/article/pii/S0021999118307125](link) - Neural-net-induced Gaussian process regression for function approximation and PDE solution - [https://www.sciencedirect.com/science/article/pii/S0021999119301032](link) - A Discussion on Solving Partial Differential Equations using Neural Networks - [https://arxiv.org/abs/1904.07200](link) ### WDM - Plasma and Warm Dense Matter Studies (LCLS) - [link](https://www-ssrl.slac.stanford.edu/lcls/talks/rlee_wdm_100400.pdf) - Frontiers and Challenges in Warm Dense Matter (collection of papers) - [link](https://www.springer.com/gp/book/9783319049113) - Fast and Universal Kohn Sham Density Functional Theory Algorithm for Warm Dense Matter to Hot Dense Plasma - [link](https://arxiv.org/abs/2004.02818) - Ab initio simulation of warm dense matter - [link](https://arxiv.org/abs/1912.09884) - Efficient formalism for warm dense matter simulations - [link](https://journals.aps.org/prb/abstract/10.1103/PhysRevB.92.161113) ### ML-DFT - Spectral neighbor analysis method for automated generation of quantum-accurate interatomic potentials - [link](https://www.sciencedirect.com/science/article/pii/S0021999114008353) - Efficient prediction of 3D electron densities using machine learning - [link](https://arxiv.org/abs/1811.06255) - Bypassing the Kohn-Sham equations with machine learning - [link](https://www.nature.com/articles/s41467-017-00839-3) - Understanding machine‐learned density functionals - [link](https://onlinelibrary.wiley.com/doi/full/10.1002/qua.25040) - Finding Density Functionals with Machine Learning - [link](https://dft.uci.edu/pubs/SRHM12.pdf) - Deep Learning and Density Functional Theory - [link](https://arxiv.org/pdf/1811.08928.pdf) - Electronic structure from conditional probabilities - [link](https://arxiv.org/abs/2007.01890) - Warm dense matter simulation via electron temperature dependent deep potential molecular dynamics - [link](https://arxiv.org/abs/1909.00231) - Completing density functional theory by machine learning hidden messages from molecules - [link](https://www.nature.com/articles/s41524-020-0310-0) ### Group Papers - Effective Static Approximation: A Fast and Reliable Tool for Warm Dense Matter Theory - [link](https://arxiv.org/abs/2008.02165) ### Misc - Studying PDEs from Mathematical Methods ..., Arfken Weber Harris - Look at NeuralPDE Julia package [link](https://github.com/SciML/NeuralPDE.jl)