Property Prediction
ML models for predicting material properties from structure — bandgap, formation energy, stability, and more.
Tutorials
matminer tutorials
Jupyter
Feature engineering
Machine Learning for Materials
Jupyter
Covers regression, classification, feature engineering for materials, and uncertainty quantification.
Regression
Feature engineering
Uncertainty quantification
Materials Informatics
Jupyter
ML for materials discovery course materials.
ML for discovery
Tools
Core Stack
pymatgen
Python Materials Genomics — analysis and manipulation
ASE
Atomic Simulation Environment
matminer
Data mining and ML for materials
JARVIS-tools
Integrated workflows for materials
emmet
Build collections of materials properties
Graph Neural Networks
MEGNet
Graph networks for molecules and crystals
CGCNN
Crystal graph networks for material properties
ALIGNN
Atomistic Line Graph Neural Network
Descriptors & ML Frameworks
DScribe
Descriptor library with various fingerprinting techniques
MAML
High-level interfaces for materials science ML
XenonPy
Material descriptors and neural network models
chemml
ML and informatics for chemical and materials data
amp
Machine-learning for atomistic calculations
Language Models
High-Throughput Frameworks
AFLOW
Ab-initio computing, high-throughput
AiiDA
Automated infrastructure for ab-initio design
atomate2
Materials science workflow library
quacc
High-throughput computational materials science
Datasets
Major Databases
Materials Project
DFT calculations, properties
500K+ materials
NOMAD
Computational materials data
19M+ calculations
AFLOW
Crystal structures, properties
3.5M+ entries
OQMD
Open Quantum Materials Database
1M+ entries
JARVIS-DFT
DFT data with ML models
40K+ materials
Benchmark Datasets
Specialized
2D Materials (C2DB)
2D materials properties
Polymer Genome
Polymers with properties
Porous Materials AI Gym
ML datasets for porous materials
Harvard OPV
Photovoltaic data with quantum-chemical calculations