This coursework tested the knowledge and understanding of information theoretic entropy with cases where the actual probability distributions were not known and estimates needed to be found. Results show the comparison of estimating the distributions using sample data naively and with Laplace smoothing. This is shown in Q5 of the CW.
The final results and answers to the questions are in the file CW1/CW1-final-answer.pdf which is supported by the code in CW1/Q5-code.py.
This coursework tested the knowledge and understanding of different forms of learning in the brain (supervised, reinforcement and unsupervised learning). This tested the ability to analyse and relate these machine learning paradigms to different parts of the brain, with the emphasis on independent reading and understanding of literature.
This coupled with an implementation of one of the previously mentioned forms of learning and an in-depth comparison to the other forms of learning. For my CW2, I implemented and studied reinforcement learning in the brain.
The implementation is available in the file CW2/reinforcement.py and the analysis is available in the file CW2/report.pdf.