Welcome to Chemically Novel Materials Discovery!

This is the documentation for mat_discover. Descending from Stochastic Clustering Variance Regression (DiSCoVeR) is a composition-based materials discovery algorithm. DiSCoVeR emphasizes exploring high-performance candidates in new chemical spaces as originally proposed by Baird, Diep, and Sparks (Digital Discovery, ChemRxiv). The library mat_discover provides a GPU-accelerated (and CPU-compatible) Python implementation and tools for analyzing the outputs. mat_discover will “suggest your next experiment” using a custom training dataset of chemical formula and target properties and a custom validation set of potential candidates. All in just a few lines of code! Get started learning about DiSCoVeR.

Star Follow @sgbaird Issue Discuss

Contents:

• DiSCoVeR
  • Citing
  • DiSCoVeR Workflow
  • Installation
  • Basic Usage
  • Bugs, Questions, and Suggestions
  • Looking for more?
• Examples
  • Reproduce Paper Results
  • Hyperparameter Combinations
  • Without Dimension Reduction
  • Real Predictions
  • Adaptive Design Comparison
  • Bare Bones
  • Cluster Plot
• Figures
  • Compound-Wise
  • Cluster Properties
  • Cluster-Wise
  • Density and Target Characteristics
  • Adaptive Design Comparison
• mat_discover API
  • Module contents
  • Submodules
  • mat_discover.mat_discover_ module
  • mat_discover.adaptive_design_ module
  • Subpackages

Indices and tables

• Index

• Module Index

• Search Page