CaltechAUTHORS
  A Caltech Library Service

3D Computer Vision Models Predict DFT-Level HOMO-LUMO Gap Energies from Force-Field-Optimized Geometries

Maser, Michael R. and Reisman, Sarah E. (2021) 3D Computer Vision Models Predict DFT-Level HOMO-LUMO Gap Energies from Force-Field-Optimized Geometries. . (Unpublished) https://resolver.caltech.edu/CaltechAUTHORS:20210721-215809226

[img] PDF - Submitted Version
Creative Commons Attribution Non-commercial No Derivatives.

637kB
[img] PDF - Supplemental Material
Creative Commons Attribution Non-commercial No Derivatives.

4MB

Use this Persistent URL to link to this item: https://resolver.caltech.edu/CaltechAUTHORS:20210721-215809226

Abstract

We investigate 3D deep learning methods for predicting quantum mechanical energies at high-theory-level accuracy from inexpensive, rapidly computed molecular geometries. Using space-filled volumetric representations (voxels), we explore the effects of radial decay from atom centers and rotational data augmentation on learnability. We test several published computer vision models for 3D shape learning, and construct our own architecture based on 3D inception networks with physically meaningful kernels. We provide a framework for further studies and propose a modeling challenge for the computer vision and molecular machine learning communities.


Item Type:Report or Paper (Discussion Paper)
Related URLs:
URLURL TypeDescription
https://doi.org/10.33774/chemrxiv-2021-11r61DOIDiscussion Paper
ORCID:
AuthorORCID
Maser, Michael R.0000-0001-7895-7804
Reisman, Sarah E.0000-0001-8244-9300
Additional Information:The content is available under CC BY NC ND 4.0 License. Fellowship support was provided by the NSF (M.R.M., Grant No. DGE-1144469). S.E.R. is a Heritage Medical Research Investigator. Financial support from the Research Corporation Cottrell Scholars Program is acknowledged. The author(s) have declared they have no conflict of interest with regard to this content. The author(s) have declared ethics committee/IRB approval is not relevant to this content.
Group:Heritage Medical Research Institute
Funders:
Funding AgencyGrant Number
NSF Graduate Research FellowshipDGE-1144469
Heritage Medical Research InstituteUNSPECIFIED
Cottrell Scholar of Research CorporationUNSPECIFIED
Subject Keywords:Machine learning; Computer vision; Computational chemistry; Delta learning
Record Number:CaltechAUTHORS:20210721-215809226
Persistent URL:https://resolver.caltech.edu/CaltechAUTHORS:20210721-215809226
Usage Policy:No commercial reproduction, distribution, display or performance rights in this work are provided.
ID Code:109962
Collection:CaltechAUTHORS
Deposited By: Tony Diaz
Deposited On:26 Jul 2021 22:46
Last Modified:26 Jul 2021 22:46

Repository Staff Only: item control page