KIST-CSRC / BespokeSynthesisPlatform

Bespoke Nanoparticle Synthesis and Chemical Knowledge Discovery Via Autonomous Experimentations

Bespoke Nanoparticle Synthesis and Chemical Knowledge Discovery Via Autonomous Experimentations

Introduction

This repository recommend optimized synthesis recipe via AI models. All of experiment equipments are controlled by Master Node, and are modularized according to the purpose of platform, such as BatchSynthesisModule, UV-VisModule, and others (might be added more platform later). Master Node contains source code of AI models, Logger, etc.

Installation

Using conda

conda env create -f requirements_conda.txt

Using pip

pip install -r requirements_pip.txt

Script architecture

Master Node
├── AI
│   └── Bayesian: Bayesian optimization
│   └── Fitness: Fitness function
│   └── SaveModel: pickle of model
└── Log
│   └── Logging_Class.py: write log for all actions in autonomous laboratory
├── Result: store data for each project
│   └── 1_Chemistry_discovery
│       └── AI_decision_process
│       └── DB
│       └── Optimization_result

Reference

1. Yoo, H. J., Kim, N., Lee, H., Kim, D., Ow, L. T. C., Nam, H., ... & Han, S. S. (2023). Bespoke Nanoparticle Synthesis and Chemical Knowledge Discovery Via Autonomous Experimentations. arXiv preprint arXiv:2309.00349.