Open Science & Resources

We put strong emphasis on engineering open source contributions and reproducibility of our papers. All of our code, datasets, and research tools are made publicly available to advance the broader research community.

Code Repositories

All code is hosted in the Lab's GitHub organization currently some of the highlights are.

Deep Chemical Reaction Networks

Code Base

code that can be used to reproduce the results presented in the paper 'Modelling chemical reaction networks using neural ordinary differential equations'.

Chemical AI Python
View on GitHub

Neural Mean Field Games

Code Base

contains four implementations of mean-field games (MFGs). The code is written in Python and all games are compatible with JAX, connecting them to the JAX ecosystem. While all games can be solved using the MFG theory, this repository offers the possibility to add a neural network to the systems of differential equations describing the games.

Mean Field Theory Chemical AI Game Theory
View on GitHub

Chemical Language Model Explainer

Code Base

This repository accompanies the paper 'Explainability Techniques for Chemical Language Models' with code to reproduce the results and apply the technique to other self-attention encoder architectures

Deep Learning Chemical AI
View on GitHub

Datasets

we have recently open sourced the dataset from our NeruIPS competition track ! .

Getting Started

Interested in using our research code or contributing? Here's how to get started:

  1. Explore the Code: Browse our GitHub organization to find projects that interest you.
  2. Read the Documentation: Each repository includes comprehensive README files with installation instructions and usage examples.
  3. Cite Our Work: If you use our code or data, please cite the corresponding publications. BibTeX entries are available on our publications page.
  4. Contribute: We welcome contributions! Open issues or pull requests on GitHub, or contact us via our contact page.