AI4C-ML: Machine Learning for Atomically Dispersed Catalysts

This repository provides machine learning workflows for atomically dispersed catalyst (ADC) design, including structure generation, performance prediction, and reaction barrier modeling.

Installation

Development Environment

Python 3.13

Validated on Linux/Windows OS

Use conda env create -f mltrain.yml to create the enviornment.

Setup

To set up the codes, run the following commands:

https://github.com/ai4cat/AI4C-ML.git
cd AI4C-ML

Repository Structure

1. Atom_Generation

Atomic structure generation module for constructing candidate ADC configurations.

• High-throughput generation of atomically dispersed catalyst structures
• Flexible control of metal centers and coordination environments
• Outputs compatible with downstream DFT and ML workflows

---

2. GPGB_AL

Machine learning pipeline for ORR performance prediction.

• GPGB (Genetic Programming + Gradient Boosting) framework
• Training and testing workflows for half-wave potential (E_1/2)
• Active learning strategy for candidate selection
• Descriptor-based modeling and feature optimization

---

3. H2O2_decom_bar

Machine learning models for reaction barrier prediction.

• Prediction of hydrogen peroxide decomposition barriers
• Supervised learning workflows for energy barrier estimation
• Supports large-scale screening of catalytic stability

---

Features

• End-to-end ML workflow for catalyst discovery
• Integration of structure generation, performance prediction, and stability evaluation
• Designed for high-throughput and data-driven catalyst screening

License

This project is licensed under the CC-BY-ND-NC License. Please see the LICENSE file for more details.