UCL MSc Machine Learning Coursework and Notes

This repository contains coursework assignments, practice notebooks, handwritten notes, and lecture notes from the MSc Machine Learning programme at University College London.

Repository Structure

Course Materials

All coursework is organized in the Courseworks/ directory:

COMP0078 - Supervised Learning

• Coursework on supervised learning methods
• Dataset: Boston housing predictions
• Topics: regression, classification, feature engineering

COMP0083 - Kernel Methods and Convex Optimisation

• ConvexOptimisation/ - Optimization coursework
• Kernel/ - Kernel methods assignment
• Topics: SVM, kernel tricks, optimization algorithms

COMP0085 - Approximate Inference

• Comprehensive coursework on probabilistic inference
• Topics: graphical models, message passing, variational inference, MCMC
• LaTeX lecture notes with extensive theoretical coverage

COMP0086 - Unsupervised Learning

• Coursework with practical datasets (MNIST, War and Peace, binary digits)
• Topics: clustering, dimensionality reduction, topic modeling
• Includes preprocessing and symbolic data files

COMP0089 - Reinforcement Learning

• 4-part coursework series (part1–part4)
• Topics: MDP, value iteration, policy gradient, Q-learning
• Hands-on RL implementation projects

COMP0171 - Bayesian Deep Learning

• coursework-one/ - Bayesian classifiers and uncertainty quantification
• coursework-two/ - Variational autoencoders with MNIST data
• Hands-on projects in probabilistic deep learning

Comprehensive Lecture Notes

Professional LaTeX formatted lecture notes for key topics:

• Approximate_Inference_Lecture_Notes.tex - 50+ pages on inference methods
• Bayesian_Deep_Learning_Lecture_Notes.tex - 15+ pages covering optimization, automatic differentiation, posterior approximation, deep architectures, and uncertainty quantification
• Kernel_Lecture_Notes.tex - Kernel methods and theory
• Reinforcement_Learning_Course_Notes.tex - RL fundamentals
• Supervised_Learning_Lecture_Notes.tex - Classification and regression
• Unsupervised_Learning_Lecture_Notes.tex - Clustering and dimensionality reduction

Supporting Materials

• Handwritten_notes/ - Scanned notes from lectures and revision
• Data files - Course datasets (MNIST, text corpora, numeric datasets)

Building the LaTeX Notes

The lecture notes are compiled using LaTeX with latexmk:

cd Notes/
latexmk -pdf <filename>.tex

All notes use consistent formatting with theorem environments, definitions, and professional layout based on lecture note best practices.

Course Progression

The curriculum progresses through:

1. Supervised Learning (COMP0078) - Foundation in regression/classification
2. Kernel Methods (COMP0083) - Advanced optimization and kernel theory
3. Approximate Inference (COMP0085) - Probabilistic graphical models
4. Unsupervised Learning (COMP0086) - Clustering and representation learning
5. Reinforcement Learning (COMP0089) - Markov decision processes and learning algorithms
6. Bayesian Deep Learning (COMP0171) - Modern deep learning with uncertainty

Technologies Used

• Python/Jupyter - Coursework implementations
• LaTeX - Professional lecture notes
• Datasets - MNIST, UCI, text corpora for practical assignments

---

These materials represent the core coursework and learning from the MSc Machine Learning programme at UCL.