🧠 Machine Learning for Chemical Engineers

Instructor: Antonio del Rio Chanona

Welcome to the Machine Learning for Chemical Engineers course! This course is designed for final-year undergraduate and master's students interested in applying data-driven and AI-based techniques to engineering problems (with an emphasis on chemical engineering).

---

🎯 Course Objectives

• Teach both fundamental and advanced machine learning (ML) and AI concepts.
• Highlight practical applications in chemical engineering.
• Intuitive explanations behind the algorithms and applications.
• Develop an understanding of what ML/AI can and cannot do.

---

Lecture 1a: Introduction to ML

• 🎥 Watch Video
• 📑 Slides

Lecture 1b: Linear regression

• 🎥 Watch Video [coming soon!]
• 📑 Slides
• 📓 Notebook: ML Models with sklearn

Lecture 1c: Maximum likelihood estimation

• 🎥 Watch Video [coming soon!]
• 📑 Slides

Lecture 2a: Neural networks

• 🎥 Watch Video [coming soon!]
• 📑 Slides

Lecture 2b: Optimization algorithms for deep learning (e.g., SGD, Adam)

• 🎥 Watch Video [coming soon!]
• 📑 Slides
• 📓 Notebook: Optimization algorithms for deep learning

Lecture 2c: Optimization in deep learning

• 🎥 Watch Video [coming soon!]
• 📑 Slides

Lecture 3: Data-driven optimization

• 🎥 Watch Video [coming soon!]

Lecture 4a: Gaussian processes

• 🎥 Watch Video [coming soon!]

Lecture 4b: Bayesian optimization

• 🎥 Watch Video [coming soon!]

Lecture 4c: Advanced Bayesian optimization

• 🎥 Watch Video [coming soon!]

Lecture 5: Unsupervised learning

• 🎥 Watch Video [coming soon!]

Lecture 6: Reinforcement learning

• 🎥 Watch Video [coming soon!]

Lecture 7a: Time-series prediction

• 🎥 Watch Video [coming soon!]

Lecture 7b: Data-driven model predictive control

• 🎥 Watch Video [coming soon!]

Lecture 8: Large Language Models (LLMs) and ChatGPT

• 🎥 Watch Video [coming soon!]

---

🗂️ Course Structure

The course consists of video lectures, slides, Jupyter notebooks, and hackathon-style coursework where you can build your own algorithm and benchmark against existing one. Each lecture includes code examples.

📚 Curriculum Highlights

1. Introduction to Machine Learning

   • What is machine learning
   • Modern ML vs expert systems ML
   • Examples of ML in engineering

2. Machine Learning Models

   • Linear regression, neural networks, decision trees
   • Gaussian processes
   • Supervised learning principles

3. Anomaly Detection

   • Probabilistic models
   • Dimensionality reduction (PCA, autoencoders)
   • Clustering methods

4. Data-Driven Optimisation

   • Bayesian optimisation
   • Derivative-free optimisation (model-based vs direct)
   • Evolutionary algorithms
   • Design of experiments & simulations

5. Data-Driven Control

   • Reinforcement learning
   • Data-driven model predictive control

6. Large Language Models (LLMs)

   • Transformers, transfer learning, self-supervised learning
   • Vocabulary and LLM pipeline
   • How does ChatGPT work?

---

🧪 Build your own algorithm

This course uses hackathon-style coursework:

There are generally 3 types of coursework mirroring ML applications for engineering:

• Data-driven optimisation: Build ML algorithms for controller tuning, design of experiments, real-time optimisation, etc.
• Unsupervised learning: Anomaly detection in the Tennessee Eastman plant.
• Data-driven control: Build a data-driven model-predictive or reinforcement learning algorithm to control chemical reactors or supply chains.

📚 All courseworks can be found here

---

📁 Repository Structure

.
├── README.md                # This file
├── syllabus.md              # (Optional) Full course syllabus
├── environment.yml          # Reproducible conda environment
├── slides/                  # PDF slides per lecture
├── notebooks/               # Jupyter notebooks per lecture/topic
│   └── lecture01_intro/
│       ├── 01_data_overview.ipynb
│       └── 02_visualization_basics.ipynb
├── data/                    # Example datasets (optional)
└── resources/               # Extra papers, readings, links