🚀 Advanced Data Science Techniques Using GPUs: A Practical Tutorial

🌟 Overview

• Event: eScience 2025
• Tutorial: Advanced Data Science Techniques Using GPUs
• Presenter: Didier Barradas Bautista
• Date: Monday September 15, 2025
• Location: Room F

🗂️ Table of Contents

1. Introduction
2. Setup Instructions
3. Tutorial Materials
4. Resources
5. Contact

🧑‍💻 Introduction

This tutorial offers a practical, hands-on exploration of advanced data science techniques powered by modern GPUs. We will dive into the NVIDIA RAPIDS ecosystem—including cuDF for fast dataframe operations, cuML for GPU-accelerated machine learning, and more—demonstrating how these tools can dramatically speed up your workflows compared to traditional CPU-based approaches.

You'll also learn how to scale your experiments and optimize models using distributed computing frameworks like Ray-tune and Optuna, as well as deep learning with Skorch. Through interactive notebooks, you'll gain practical experience in:

• Accelerating data preprocessing and analysis with GPUs
• Training and tuning machine learning models at scale
• Leveraging distributed hyperparameter optimization
• Integrating RAPIDS with popular Python data science libraries

Whether you're new to GPU computing or looking to take your data science skills to the next level, this tutorial will equip you with the knowledge and tools to harness the full power of GPUs for your research and projects. Get ready to supercharge your data science journey!

⚙️ Setup Instructions

Using Google Colab

1. Click the Open in Colab badge next to each notebook above, or open the notebook link directly in Google Colab.
2. Sign in with your Google account if prompted.
3. Go to the menu: Runtime > Change runtime type.
4. In the "Hardware accelerator" dropdown, select GPU and click Save.
5. If the notebook includes a setup cell (for RAPIDS, Optuna, etc.), run it first before running other cells.
6. To save your work, go to File > Save a copy in Drive.

Tip: If you see any errors about missing packages, re-run the setup cell or restart the runtime (Runtime > Restart runtime).

Instructions for setting up your environment locally (RAPIDS, Optuna, etc.) will be provided soon.

📂 Tutorial Materials

• Hands-on notebooks
  • Notebook 1: Hyper-parameter Optimization with Sklearn + RAPIDS
  • Notebook 2: Hyper-parameter Optimization with RayTune + RAPIDS
  • Notebook 3: Hyper-parameter Optimization with Skorch + RAPIDS
• Code samples for GPU-accelerated workflows in this repository here GitHub Repository

🔗 Resources

• eScience 2025 Conference Website
• NVIDIA RAPIDS
• Skorch
• Optuna
• Ray
• GitHub Pages Documentation

📬 Contact

For questions, contact Tutorial Organizer.