Data Science & Machine Learning — Master’s in Physics (PHYMSCFUN02)

This repository hosts all teaching materials, notebooks, dataset pointers, and reproducible workflows for the course “Data Science and Machine Learning” (Master’s in Physics with a focus on Fundamental Physics, second semester 2025). It follows the official syllabus covering statistics, time series & image analysis, Fourier/Monte Carlo methods, modern ML (CNNs, AEs, VAEs/GANs), and physics‑informed neural networks—with strong emphasis on open, reproducible science.

---

🎯 Course Goals & Learning Outcomes

By the end of the course, students will be able to:

• Apply advanced statistical methods to analyze scientific data and solve physics problems.
• Work fluently with Python scientific libraries for time series and scientific images.
• Use Fourier analysis and Monte Carlo techniques in physical modeling.
• Understand and implement AI/ML approaches (supervised/unsupervised, CNNs, autoencoders, generative models).
• Integrate physics + ML concepts into research‑grade projects and communicate results in technical reports/papers.

---

🧭 Course Structure (Units & Timeline)

Unit 1 (Weeks 1–3): Dimensionality reduction & time series

• PCA, SVD, t‑SNE, UMAP, autoencoders; feature extraction; Fourier/wavelets; ARIMA & autocorrelation.

Unit 2 (Weeks 4–8): Simulation & core ML

• Monte Carlo (random walks, diffusion), supervised/unsupervised learning, cross‑validation, CNNs, AEs for images.

Unit 3 (Weeks 9–13): Advanced models & applications

• PINNs for PDEs with boundary conditions, scientific image segmentation, VAE/GAN, transfer learning.

Unit 4 (Weeks 14–16): Scientific communication & final project

• Paper structure & writing, reproducibility (Overleaf/GitHub), manuscript prep, presentations & defense.

---

📝 Assessment & Deliverables

• Weekly participation & formative tasks: 20%
  • Class participation (10%), mini technical reports & autonomous tasks (10%).
• Midterm (Week 8): 30%
  • Midterm report (20%) + partial results presentation (10%).
• Final (Weeks 14–16): 50%
  • Final project: draft of a scientific paper (30%) + final presentation & defense (20%).
• Recovery follows graduate regulations if final grade < 7.5 (at instructor’s discretion).

---

🧪 Repository Layout

.
├── week01/
│   ├── lecture_intro.pdf
│
├── week02/
│   ├── lecture.pdf
│   ├── hands_on.ipynb
│   ├── hands_on_solution.ipynb
│   └── autonomous_activity.ipynb
│
├── ...
│
├── week16/
│   ├── lecture.pdf
│   ├── hands_on.ipynb
│   ├── hands_on_solution.ipynb
│   └── autonomous_activity.ipynb
│
├── projects/
│   ├── README.md
│   ├── project01/
│   │   ├── data/
│   │   ├── src/
│   │   └── notebooks/
│   ├── project02/
│   │   ├── data/
│   │   ├── src/
│   │   └── notebooks/
│   ├── ...
│   └── project05/
│       ├── data/
│       ├── src/
│       └── notebooks/
│
├── env/
│   ├── environment.yml
│   └── requirements.txt
│
├── CITATION.cff
├── LICENSE
└── README.md

Tip: Large data should be stored via Git LFS or referenced by download scripts (see data/README.md).

---

⚙️ Environment & Setup

Option A — Conda (local)

# Create environment
conda create -n mfff-ml python=3.10 -y
conda activate mfff-ml

# Core scientific stack
pip install -r env/requirements.txt

env/requirements.txt (example)

numpy
scipy
pandas
matplotlib
seaborn
scikit-learn
umap-learn
pywavelets
sympy
jupyter
ipykernel

Option B — Google Colab

Open notebooks in Colab; GPU is optional but recommended for CNNs/AEs/VAEs/GANs. Ensure the correct runtime (GPU) under Runtime → Change runtime type.

---

▶️ How to Run

# 1) Activate environment
conda activate mfff-ml

# 2) Launch Jupyter
jupyter lab  # or jupyter notebook

# 3) Open notebooks under notebooks/02_dimensionality_reduction and run cells

🔬 Reproducibility & Good Practices

• Use GitHub Issues for task tracking; each PR must include:
  • Clear description, checklist, and links to related notebook cells.
  • Reproducibility note (random seeds, environment hash).
• Use Overleaf for writing the midterm/final manuscripts and keep a synced PDF under reports/.
• Document data provenance (source, license), and set random seeds for stochastic methods (t‑SNE/UMAP/NNs).

---

👥 Collaboration & Office Hours

• Teamwork is expected; cite all shared components clearly in notebooks and reports.
• Office hours: Monday–Thursday, 19:00–20:00 (local).

---

🤖 Academic Integrity & AI Use

• AI tools may assist learning (e.g., coding hints) only with explicit attribution.
• For evaluations (exams/reports/projects), generating text/code/results with AI without proper attribution is a serious offense. Maintain originality, transparency, and data integrity.

---

🧾 Licensing & Citation

• Code is released under the repository’s LICENSE (see file).
• If you use these materials in academic work, please cite the course and include a link to this repository.
• Course alignment and official information are based on the PHYMSCFUN02 syllabus.

---

Instructor: Dr. Hernán Andrés Morales‑Navarrete
Program: Master’s in Physics with a focus on Fundamental Physics — Second Semester 2025.