Teaching Statistics and Mathematical Modelling tutorials for first year Biomedical Sciences undergraduate students.

Project title: Complex Biological Synapses for Unsupervised Learning in Non-Stationary Environments. Biological synapses exhibit plasticity dynamics on multiple timescales, which is hard to model accurately using the simplified representation in the form of a single weight value. The aim of the project is to utilise complex synapse models to achieve better unsupervised learning performance when the learning task or environment is non-stationary. The model proposed by Benna and Fusi (2016) is… Show more Project title: Complex Biological Synapses for Unsupervised Learning in Non-Stationary Environments. Biological synapses exhibit plasticity dynamics on multiple timescales, which is hard to model accurately using the simplified representation in the form of a single weight value. The aim of the project is to utilise complex synapse models to achieve better unsupervised learning performance when the learning task or environment is non-stationary. The model proposed by Benna and Fusi (2016) is used as a starting point, and potentially there will be deviation from this, e.g. by using gradient based approaches to optimise the model of complex synapses. Show less

Connected Care, Philips Research China.Natural Language Processing for clinical data: utilised deep learning architectures including the Transformer and the Denoising Autoencoder to standardise medical terms in electronic medical records.Inventory forecasting for online retailers: produced code for generating business report and model performance assessments from inventory data and output of AI models.Coded with Python, heavily using Keras and pandas.

Department of Applied Mathematics and Theoretical Physics, University of Cambridge.Supervisor: Dr Stephen Eglen.Worked on neuroscience examples for CODE CHECK platform (https://codecheck.org.uk/).