Giuseppe La Tona, Ph.D.

I am a Researcher with the National Research Council of Italy (CNR), Institute of Marine Engineering (INM). My research interests include the development of Energy Management Systems and of Forecasting Models.

2 minute read

I’m excited to announce that I’ll be co-presenting Tutorial #4 at IECON 2025 in Madrid, Spain!

Tutorial Details

Title: Deep Learning-based Forecasting of Energy Related Time Series: Tuning, Evaluation, and Reproducibility

Lecturers:

Conference: IECON 2025 - 51st Annual Conference of the IEEE Industrial Electronics Society
Date: October 14, 2025 | 9:30 AM - 12:30 PM
Location: Commendador Room, Madrid, Spain

About the Tutorial

This tutorial provides an informative and dedicated training session on deep learning-based forecasting methods for energy-related time series. The session covers fundamentals and state-of-the-art techniques in:

  • 🧠 Deep learning architectures for time series forecasting
  • 🎯 Hyperparameter tuning strategies
  • 📊 Model evaluation methodologies
  • 🔬 Reproducibility best practices in energy forecasting

Hands-On Notebooks

The tutorial includes practical Jupyter notebooks covering:

  1. Environment Setup & Verification - Verify all dependencies are correctly installed
  2. Data Exploration - Exploratory data analysis of household power consumption
  3. Preprocessing & Feature Engineering - Data preprocessing, windowing, and feature engineering
  4. Forecasting Models - Deep learning models for energy forecasting (baselines, feedforward networks, LSTM)
  5. Figure Preparation - Generate presentation-quality figures

All notebooks are designed to run seamlessly in:

  • Local Development using DevContainers or Poetry
  • Google Colab for easy access without local setup

Resources

📂 Tutorial Materials

All tutorial materials, including code, notebooks, and datasets, are available on GitHub:

🔗 GitHub Repository: iecon2025_tutorial

📊 Presentation Slides

Download: 📥 PDF slides

Interactive version available on the tutorial GitHub repository

Technical Stack

The tutorial uses modern Python tools and libraries:

  • Python 3.12 with Poetry for dependency management
  • TensorFlow 2.19.0 and Keras 3.10.0 for deep learning
  • NumPy, Pandas, SciPy for data manipulation
  • scikit-learn for machine learning utilities
  • matplotlib, seaborn, plotly for visualization
  • DevContainer support for consistent development environments

All dependencies match Google Colab versions (as of October 2025) for maximum compatibility.

Why This Tutorial Matters

Energy forecasting is crucial for:

  • ⚡ Smart grid management and optimization
  • 🏠 Building energy management systems
  • 🚢 Shipboard electrical microgrid control
  • 🌱 Renewable energy integration
  • 💰 Energy trading and demand response

Deep learning methods have shown remarkable success in capturing complex temporal patterns in energy consumption data, but proper training, evaluation, and reproducibility are essential for real-world deployment.

For More Information

📍 For IECON 2025 Attendees

When: Tuesday, October 14, 2025
Time: 9:30 AM - 12:30 PM
Room: Commendador Room

Looking forward to seeing you in Madrid! Don’t miss this hands-on session on deep learning for energy forecasting.

#IECON2025 #DeepLearning #EnergyForecasting #MachineLearning #TimeSeriesAnalysis

You May Also Enjoy