Machine Learning Projects
Thought this would be a nice way to visualise my projects.
This is a terrain I made in Unity, the height-map is a smooth approximation to the actual Mariana trench. I show different optimisation algorithms searching the domain, trying to minimise the height. The algorithms in the video: ABC (Artificial Bee Colony), Gradient Descent, Adam (ADAptive Momentum estimation), Gibbs Sampling and Metropolis Hastings.
Reinforcement Learning Projects
- Implementation of Richard Sutton & Andrew Barto's Intro to RL
- Race car Reinforcement Learning agent in Unity
Coming Soon - Unity ML agents project
Coming Soon
- RL is in my opinion the most exciting area of machine learning, naturally encompassing Supervised and Unsupervised techniques within it. I've explored a few interesting areas and I will continue to add to this section.
Supervised Learning Projects
- Sequence Models
- Neural ODEs
Coming Soon
- Supervised Learning is probably responsible for the majority of the big successes in ML. The main project I have in this page is on sequence models. I include the sequence algorithms in "Deep Learning" Goodfellow, Bengio & Courville. A Neural Turing Machine, Graves et al. Attention Mechanisms, such as location-based, Bahdanau, self and multi-head. A Transformer, Vaswani et al.
Unsupervised Learning Projects
- Variational Autoencoder
Coming Soon