Machine learning for chemical discovery (2020-08-17T00:00:00.000000Z)

TL;DR

Comment on recent breakthroughs in this emerging field of novel machine learning tools to obtain chemical knowledge from curated datasets, and discuss the challenges for the years to come.

Abstract

Discovering chemicals with desired attributes is a long and painstaking process. Curated datasets containing reliable quantum-mechanical properties for millions of molecules are becoming increasingly available. The development of novel machine learning tools to obtain chemical knowledge from these datasets has the potential to revolutionize the process of chemical discovery. Here, I comment on recent breakthroughs in this emerging field and discuss the challenges for the years to come.

Authors

A. Tkatchenko

10 papers

Machine learning for chemical discovery

TL;DR

Abstract

Authors

References26 items

QM7-X: A comprehensive dataset of quantum-mechanical properties spanning the chemical space of small organic molecules

Accurate Many-Body Repulsive Potentials for Density-Functional Tight Binding from Deep Tensor Neural Networks.

A Deep Learning Approach to Antibiotic Discovery

Exploring chemical compound space with quantum-based machine learning

Machine learning for molecular simulation

The ANI-1ccx and ANI-1x data sets, coupled-cluster and density functional theory properties for molecules

Learning to Use the Force: Fitting Repulsive Potentials in Density-Functional Tight-Binding with Gaussian Process Regression.

Accurate and transferable multitask prediction of chemical properties with an atoms-in-molecules neural network

Accurate molecular polarizabilities with coupled cluster theory and machine learning

Machine learning for molecular and materials science

Quantum-Mechanical Relation between Atomic Dipole Polarizability and the van der Waals Radius.

Realistic Atomistic Structure of Amorphous Silicon from Machine-Learning-Driven Molecular Dynamics.

Towards exact molecular dynamics simulations with machine-learned force fields

Planning chemical syntheses with deep neural networks and symbolic AI

Predicting protein–ligand affinity with a random matrix framework

Design of efficient molecular organic light-emitting diodes by a high-throughput virtual screening and experimental approach.

Quantum-chemical insights from deep tensor neural networks

The chemical space project.

Computational design of molecules for an all-quinone redox flow battery

Quantum chemistry structures and properties of 134 kilo molecules

The high-throughput highway to computational materials design.

Gaussian approximation potentials: the accuracy of quantum mechanics, without the electrons.

Reprints and permission information

Supplementary Materials for Machine learning unifies the modeling of materials and molecules

ible neural network potential with DFT accuracy at force fi eld computational cost †

Chemical space

Field of Study

Journal Information

Name

Volume

Venue Information

Name

Type

URL

Alternate Names