'layman's' example

[vanAmsterdam]

Thanks for providing this code and shiny app along with your article in Nature Machine Intelligence.
I'm greatly interested in the interplay between causality and machine learning, and personally work on medical applications. As an ML practitioner working on applications your method seems interesting, but the provided examples are not very accessible.

Could you provide an example that may relate to real-world data? I'm thinking:

• clinical parameters from patients, generated by distinct disease phenotypes (e.g. records of age, sex, blood pressure, lung function etc, simulated according to some disease model with different underlying factors)
• images (e.g. generated through different causal models)

I'd happily think along with generating some example simulations

[andandandand]

Hi @vanAmsterdam! The article on Nature Machine Intelligence refers to the code on this repository https://github.com/allgebrist/Causal-Deconvolution-of-Networks

[andandandand]

@vanAmsterdam if you'd like to collaborate on a project using the block decomposition method for medical images please send me an email at [email protected] and take a look at this repository https://github.com/andandandand/ImageAnalysisWithAlgorithmicInformation