Deep neural network solution of the electronic Schrödinger equation · HackerPapers discussions

Deep neural network solution of the electronic Schrödinger equation — mentioned by currymj

Tue, 21 Jul 2020 16:27:22 GMT

https://arxiv.org/abs/1909.08423 and https://arxiv.org/abs/1909.02487 are some examples I've been looking at recently.

Paper and discussion history. A mention is not an endorsement.

Deep neural network solution of the electronic Schrödinger equation — mentioned by currymj

Tue, 21 Jul 2020 16:24:23 GMT

this is very cool. I was thinking specifically of this and related approaches https://arxiv.org/abs/1909.08423 where they search for the ground state by iteratively using an MCMC sampler and doing SGD. The innovation is a network architecture that takes classic approaches from physics and judiciously replaces parts with flexible NNs. I had not even considered how things might work if you actually want to think about time. Do you know if anybody has been running this NN+DiffEq solver stuff on big HPC systems that also have GPUs? If you know of any papers where they tried this, would be…

Paper and discussion history. A mention is not an endorsement.

Deep neural network solution of the electronic Schrödinger equation — mentioned by jhrmnn

Fri, 22 Nov 2019 20:51:11 GMT

This year, I’ve been working on representing electronic wave functions with neural networks: https://arxiv.org/abs/1909.08423 I come from a physics background, and my impression is that NNs are a cool computational tool that every computational physicist probably should have in their toolbox (at least having a sense of their capability). But so far I haven’t seen any truly revolutionary advance with NNs acrose the whole of materials science.

Paper and discussion history. A mention is not an endorsement.