fastHDMI - Fast High-Dimensional Mutual Information Estimation

Kai Yang

Contact: <kai.yang2 "at" mail.mcgill.ca>

GPG Public Key Fingerprint: B9F863A56220DBD56B91C3E835022A1A5941D810

Welcome to the fastHDMI repository! This repository contains the source code and results for the fastHDMI package, as detailed in my paper: fastHDMI: Fast Mutual Information Estimation for High-Dimensional Data.

Package Information

• Package Manual: The detailed manual for the fastHDMI package is available here.
• PyPI Package: You can find the published package on PyPI here.

ABIDE Data Analysis

• Data: This repository includes an analysis of the (pre-processed) ABIDE data.
• Jupyter Notebook: The analysis is summarized in this Jupyter notebook. Running the notebook will generate the necessary Python and bash scripts to utilize the fastHDMI package for data analysis.
• Server Execution: These scripts are configured to run on a server, specifically Compute Canada in this case. After executing the scripts on the server, rerun the Jupyter notebook with the returned data files (e.g., .npy files) to produce the plots and results presented in the paper. The output plots are in PDF format.

Computational Resources

• Resource Usage: The seff-[jobID].out files provide a summary of the computational resources used for each job, generated by the seff [jobID] command.
• Compute Canada Documentation: More information about running jobs on Compute Canada can be found here.

For further inquiries or issues, please contact me at <kai.yang2 "at" mail.mcgill.ca>.