MLT_LR

Learning route of my Machine Learning Theory course at UNAL 2025

Learning path in Machine Learning Theory and Practical Development (UNAL 2025)

---

Description

This repository documents my learning route in Machine Learning Theory (MLT),
including:

• Notebooks and experiments for practical model implementation.
• Optimization workflows (Bayesian, Optuna, skopt).
• Theoretical material: books, notes, and references.
• Development environment to run and reproduce results.
• Cloned repositories for study, testing, and integration.

---

Branch Strategy

To keep the work organized, this repo uses a branching model:

• main
  Main branch. Contains stable, reviewed, and clean code/documentation.
  Always reflects the state of the project that is ready to be delivered or referenced.

• dev
  Development branch. Contains ongoing work and integrates features before merging into main.
  All new features or experiments should be branched off dev.

---

Repository Structure

MLT_LR/
├── .gitignore # Git ignored files (at repo root)
├── README.md # Main repository description (at repo root)
│
└── mltEnv/ # Python virtual environment + main repo structure
│ ├── Include/ # venv internal
│ ├── Lib/ # venv internal
│ ├── Scripts/ # venv internal
│ ├── README.md # Readme to describe development environment
│
│ ├── cloneRepo/ # Cloned repositories from professors or external projects
│ │ └── README.md
│
│ ├── content/ # Study materials and resources on ML theory
│ │ ├── books/ # Books and references
│ │ └── README.md
│ │
│ ├── work/ # Includes organized submissions, project ,reports, and deliverables
│ │ └── README.md
│ │
└── requirements.txt # Dependencies for reproducibility (on a future after some model implementation )

how to RUN nbs

1. Set up the environment (example with Python venv)

# Create virtual environment
python -m venv mltEnv

# Activate it
source mltEnv/bin/activate     # Linux/Mac
.\mltEnv\Scripts\activate      # Windows PowerShell

# Install dependencies
pip install -r requirements.txt

Then ,navigate to the correct folder depending on task: -DevContent/ → Run and develop Jupyter notebooks or ML scripts. -content/ → Access theory, books, and lecture notes. -work/ → View reports, deliverables, and documentation. -cloneRepo/ → external repositories.

In order to run notebooks,

From the repo root : jupyter notebook work/

or open the .ipynb files directly in VSCode with the Python extension, also you can download and import on colab or any other environment.

---

Projects and Experiments

• Algebra and probabilistic analysis as a foundation for deeper comprehension of model architectures.
• Gradient Descent and regressors approaches for linear models.
• Bayesian Optimization for kernel ridge regression.
• Housing datasets (California, etc.) with analytical + iterative methods.
• NFL Big data bowl 2026 to predict player movement while the ball is in the air.

---

Notes and Resources to develop

All inside PYDevEnv/:

• Books & class PDFs → content/ss'sBooks/
• Class notes & theory → content/lectures/
• work class → work/

---

Roadmap

• Create README base
• Add repo structure
• Upload repo from course
• Document first experiments and tasks (linear regression, Bayesian optimization)
• First partial test (ML on NFL DB)
• Deliver work as:
  • assigned tasks
  • mathematical/model demonstrations
• second test (DL on NFL DB)
• Project

Author

Mateo Almeida (Macreat)

Course: Machine Learning Theory @ UNAL 2025

GitHub: @Macreat

License

License: no

Usage: Academic and learning purposes