AI-Projects — ML Theory & Optimization Notes

This repository is a curated set of theory-heavy AI / ML projects (math + optimization), mostly documented through PDF reports and companion code when relevant.

It is intentionally focused on fundamentals: understanding the algorithms, not just using black boxes.

For applied / full projects (e.g., Deep RL racing), see my other repositories.

---

Structure

AI-Projects/
├── ADAM-MLP-Implementation/
├── SVM - Support Vector Machine/
├── Levenberg-Marquardt/
├── Mont-Carlo-Algorithms/
└── README.md

---

Contents

ADAM vs SGD — MLP from scratch

• Manual MLP implementation (forward + backprop) with Adam optimizer
• Comparison against SGD on controlled experiments → theory/ADAM-MLP-Implementation/

Support Vector Machines (SVM) — low-level study

• Step-by-step implementations (linear + kernel methods)
• Progressive study: primal/dual, soft margin, Gaussian & polynomial kernels
• Hyperparameter impact analysis (C, σ, kernel degree) → theory/SVM - Support Vector Machine/

Levenberg–Marquardt

• Nonlinear regression via Levenberg–Marquardt
• Comparison with alternative optimization strategies → theory/Levenberg-Marquardt/

Monte Carlo / Quasi-Monte Carlo

• Numerical integration & sampling-based optimization
• QMC methods, clustering effects, and related experiments → theory/Mont-Carlo-Algorithms/

---

Most folders include a PDF report (primary reference). This repo is kept intentionally compact and focused on fundamentals.