Skip to content
/ UCL Public

Coursework from the MSc Machine Learning at University College London, 2021/2022.

License

Notifications You must be signed in to change notification settings

fededagos/UCL

Repository files navigation

UCL

This repository contains a collection of coursework from my MSc in Machine Learning at UCL (2021/2022). Content of the various folders ranges from reports, paper reviews, coding exercises and end-to-end projects.

Below is a summary of the various topics tackled in the various projects present here.

Supervised Learning

Kernelisation of algorithms, sample complexity analysis, study of generalisation behaviour of common algorithms (perceptron, winnow, knn,...). Kernel perceptron from-scratch implementation and analysis on MNIST.

Statistical Natural Language Processing

Group project on argument mining (argument component identification to be precise). The full codebase for the project can be found here.

Machine Learning Seminar

Review of Deep Kernel Learning. Group presentation on Neural Importance Sampling

Reinforcement learning

Miscellaneous coursework covering the entire module, from bandit algorithms to Deep Reinforcement learning using JAX and haiku.

Graphical models

Summative coursework on Bayesian Belief Networks, inference on graphs, LDPC codes and more.

Bayesian Deep learning

Bayesian classifiers and posterior approximation. Variational Autoencoders. Uncertainty quantification and calibration in pre-trained Deep Neural Networks.

Applied Machine learning

From-scratch implementation of CART trees, random forests, ADAboost, Support Vector Machines.

About

Coursework from the MSc Machine Learning at University College London, 2021/2022.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages