Welcome to the repository for "Computational Reproducibility in Machine Learning: A Hands-On Workshop." This repository contains files that facilitate the hands-on exercises discussed in the workshop slides, along with the workshop slides and a setup guide for configuring development environments on Windows, macOS, and Linux.
Computational reproducibility, defined as the ability to consistently recreate the results of a computational analysis using the same data, code, and methods, is a cornerstone of reliable and impactful research. This workshop explores the principles of computational reproducibility and demonstrates how to integrate practical tools into research workflows to enhance transparency, reliability, and efficiency.
The first version of this workshop was given at:
Data Science in Multi-Messenger Astrophysics Program
University of Minnesota
February 25, 2025
This workshop is based on the paper:
[1] J. Shenouda and W. U. Bajwa, “A guide to computational reproducibility in signal processing and machine learning,” IEEE Signal Processing Magazine, vol. 40, no. 2, pp. 141–151, Mar. 2023, doi: 10.1109/MSP.2022.3217659.
This repository includes:
- Pre-Workshop-Setup-Guide.md (and PDF version): Instructions for setting up the required development environment.
- 2025-02-25_UMN-Workshop-Slides.pdf: The presentation slides used during the University of Minnesota workshop.
- Files supporting hands-on exercises for Git, refactoring, virtual environments, Docker, and random number generation.
Each directory contains files corresponding to a specific topic covered in the workshop:
conda-env-repro-workshop/ # Virtual environments and dependency management
docker-repro-workshop/ # Containerization using Docker
git-repro-workshop/ # Version control with Git
refactoring-repro-workshop/ # Code refactoring techniques
rng-repro-workshop/ # Random number generation and reproducibility
2025-02-25_UMN-Workshop-Slides.pdf # University of Minnesota workshop slides
Pre-Workshop-Setup-Guide.md # Setup instructions
Pre-Workshop-Setup-Guide.pdf # PDF version of setup guide
README.md # This file
This directory contains two shell scripts for testing Git-based version control concepts:
git-create-initial-files.sh: Generates an initial set of files to track with Git.git-create-additional-files.sh: Adds another set of files, allowing users to practice tracking new additions and understanding Git history.
This repository is shared under the Creative Commons Attribution 4.0 International (CC BY 4.0) license. You are free to share and adapt the material as long as you provide appropriate credit. For details, see the full license text at: https://creativecommons.org/licenses/by/4.0/