Daniel Catchpole
Hi, I'm currently a first year PhD student in the MathSys II CDT. My current research is surrounding the use of multimodal data in creating and improving new machine learning methods and models in classification problems. As varying types of multimodal data exist in a variety of contexts, I'm interested in an equally wide variety of applications, however I am currently focusing primarily on healthcare applications.
Past Projects
MSc Individual Project (Summer 2020) - Using the structure of the body with graph convolutional neural networks for human activity recognition
Investigated the suitability of Graph Convolutional Neural Networks (GCNs) in Human Activity Recognition (HAR) problems, especially when data is recorded in a multimodal, multisensory environment. Supervised by Dr Tanaya Guha and Dr Victor Sanchez.
MSc Group Project (2020) - Using Generative Adversarial Networks to create multi-channel images of cells undergoing macropinocytosis
Evaluated various generative architectures to learn the mapping between a reference protein marker and several secondary proteins within a cell, enabling us to map the spatio-temporal order of actin recruitment within it. Supervised by Dr Till Bretschneider with data from 3i.
Teaching
This Year
CS355 - Digital Forensics (Term 2, January - March 2021), Teaching Assistant
CS342 - Machine Learning (Term 1, October - December 2020), Teaching Assistant
Past
CS313 - Mobile Robotics (Term 1, October - December 2019), Teaching Assistant
CS141 - Functional Programming (Term 2, January - March 2019), Teaching Assistant
Contact
Office D2.05, Complexity Science, Zeeman Building