Skip to content

Latest commit

 

History

History
20 lines (15 loc) · 876 Bytes

README.md

File metadata and controls

20 lines (15 loc) · 876 Bytes

Myerson Values to Explain Cheminformatics Data

myerson -- A package to explain GNNs

We developed a python package to calculate Myerson values and use them as explanations for GNNs, which can be downloaded using pip or conda / mamba:

pip install myerson[explain]
conda install myerson # doesn't include PyTorch, PyG dependencies

Explaining Cheminformatics Data

This repository contains code we used to validate the us of Myerson values as XAI method for GNNs in the context of cheminformatics. For the proof-of-concept experiments include the data and code, while for the experments on the dataset of kinase inhibitors we are not free to share the data.

Note that this repository is a combination of several private repositories and that, while we tried to resolve any dead paths, it is possible that we missed something.

Citation

TBD.