If you are a Warwick student looking for a 3rd/4th year or research project, feel free to reach out to me if any of the following projects interest you!

What will you get by working on these projects?

Most of my research is in the domain of machine learning / artificial intelligence for biomedical applications. Working with me would give you a chance to develop expertise in machine learning, image analysis and the ability to make a difference in the world by working on healthcare challenges! You may be able to publish your results enabling you to apply for further academic or industrial research in Artificial Intelligence.

Below I list different projects that I am interested in. For further details do check my webpage (and the videos and research papers there): https://warwick.ac.uk/fac/sci/dcs/people/fayyaz_minhas/

I want to discuss some of these or other project ideas. How can I reach out?

If you would like to discuss one of the following projects or a project in a related domain, you are welcome to email me. Given the large volume of emails, my response can be slow. Feel free to send reminders if I havent responded to you within 3-4 days. I typically schedule regular project meetings to discuss and advise on project options once there are a sufficient number of students who are interested in different projects.

My Current Remaining Capacity: Not taking any students at the time

Bioinformatics Projects

Are you interested in using computational science esp. AI and machine learning in making the next generation of medicine or advancing our understanding of the underlying biological causes of diseases? If so, please read on.

Note: No knowledge of biology is required for starting these projects.

Keywords: Machine learning and AI

Required Skills: Python Programming, Machine Learning, Image Analysis

Strongly recommended: Before starting the projects, bring yourself upto speed on ML via my data mining and machine learning moduleLink opens in a new window.

BI-1: Developing machine learning models for prediction of protein-compound interactions

Goal: How does caeffine keep you awake? It binds to certain proteins in your body. Identifying protein-compound interactions like this lies at the core of drug design. In this project we shall explore machine learning methods for this purpose.

Status: Available

BI-2: Machine learning models for countering antimicrobial resistance

Goal: Antimicrobial resistance is a global threat. In this project we shall explore how machine learning can be used to coutner it.

Status: Available

Computational Pathology Projects

Are you interested in developing machine learning models to help diagnose and treat cancer? The goal of the following projects is to develop effective solutions to problems in computational pathology through machine learning.

Note: No knowledge of biology or medicine is required for starting these projects.

Keywords: Machine learning and AI

Required Skills: Python Programming, Machine Learning, Image Analysis

Strongly recommended: Before starting the projects, bring yourself upto speed on ML via my data mining and machine learning moduleLink opens in a new window.

Project listing

CP-0: Predicting disease trajectories from laboratory measurements

Goal: Is it possible to predict the risk of developing a certain disease in the future from past measurements of a person's lab measurements? In this project, we will be developing machine learning methods for this purpose.

Status: Available

CP-1: Graph Neural Network Based Predictive Modelling of Computational Pathology Data

Goal: Building on our existing and ongoing work in the domain of graph neural networks for computational pathology, the goal of this project is to use graph based techniques for capturing how cells in biological samples or tumors interact in the development of diseases such as cancer.

Status: Available

CP-2: Robustness evaluation and critical assessment of deep learning models for Computational Pathology

Goal: Can you trust the output of machine learning models for diagnosing cancer? Is your predictor of cancer diagnosis better than this other one? In this project, we shall develop analysis pipelines to compare different machine learning models for computational pathology and develop tools to make sure that machine learning models are robust enough to be trusted in clinical settings.

Status: Available

CP-3: Application of large computer vision models to clinical problems in computational pathology

Goal: In recent years, there has been rapid development of large vision models such as the Segment Anything ModelLink opens in a new window (SAM) and DINOLink opens in a new window. These hold great potential for clinical computational pathology problems which will be explored in this project together with integrating these into computational pathology workflows with TIA Toolbox or QuPath.

Status: Available

CP-4: Image registration techniques in Computational Pathology

Goal: Cancer diagnosis typically involves different images of tissue. However, to get interesting insights into disease mechanisms, these images need to be aligned or "registered". In this work we will develop methods for this purpose.

Status: Available

CP-5: Neural Compression Techniques for Computational Pathology Applications

Goal: Cancer images can be quite large and it is important to compress them for storage and transfer. In this work, we will develop strategies for compressing these images with neural network inspired schemes.

Status: Available

CP-8: Modelling spatial heterogeneity of tumors

Goal: The composition of in tumors varies significanlty. In this work, we will develop methods that can capture spatial variability of tumors through machine learning approaches.

Status: Available

CP-9: Developing explainable models for computational pathology

Goal: Can an AI model explain what it is doing when predicting cancer in term of what features are important? In this work, we will develop such models for computational pathology.

Status: Available