About

I am Chris. I am a Post-Doctoral Reseacher at the Cancer Research UK Scotland Institue in Glasgow, Scotland. I have a passion for artificial intelligence, machine learning and how they can be applied to solve the most intractable problems we face today.

Current Research

I am currently researching novel methods in self-supervised deep learning for whole slide representation learning, focusing in the context of pathology for cancer diagnosis. My research focuses on leveraging self-supervised techniques to learn from vast amounts of unannotated histopathological images at both the tile and cell levels. By developing models capable of understanding and representing complex patterns in these images, I aim to enhance the accuracy and efficiency of cancer diagnosis. This approach reduces the dependency on extensive manual annotations and enables the discovery of novel insights from histological slides, with the end going being to contribute to improved patient outcomes in cancer care.

Background

My foundational training is in A.I., where I hold a 1st Class Degree in Computer Science with Artificial Intelligence from Heriot Watt University in Edinburgh. During my time there, I completed projects in areas like recommender systems, A.I. planning, computer vision, machine learning and neural networks. In my final year, my dissertation project was on the effectiveness and optimisation of Convolutional Neural Networks (CNNs) for the classification of Lymphoma biopsies. I built the networks using Tensorflow V1 and Keras and wrote a custom architecture for evolving CNNs using genetic algorithms in Python3. Full Text and code.

Subsequently, I contributed to the GRID project working on the natural language A.I. called Alana to develop and integrate an intelligent agent into a smart building. Here I mainly worked with Python, RASA NLU and docker to extend the proprietary Alana software.

I hold a PhD in deep learning and cancer science from the University of Glasgow, obtained in 2024. My doctoral research focused on developing novel methods for tumour bud scoring in Colorectal Cancer. My research included creating custom generative adversarial networks (GANs) for stain-to-stain translation from Hematoxylin and Eosin (H&E) to Immunohistochemistry (IHC), incorporating innovative methods of image registration and alignment. I also developed bespoke datasets and a custom segmentation network to detect tumour buds directly in H&E images. Additionally, I designed a self-supervised model to learn histological features, determine the invasive margin, and interpret the resulting embeddings using aggregation networks to automatically assess aggression in colorectal cancer.

Extra-Curricular

Outside of work, I enjoy a variety of hobbies. Programming is a passion that extends beyond my research, and I love experimenting with new technologies in my spare time. Hiking is a favorite way to relax and stay active, offering a great break from the lab. I also have a keen interest in photography, capturing moments from my hikes and travels. Speaking of travel, I love exploring new places and experiencing different cultures, which always provides fresh perspectives and inspiration. I am also a keen reader, with a particular interest in science fiction novels. I am always on the lookout for new recommendations, so feel free to share your favorites with me!