Curriculum Vitae - Matthew Cockayne
For specific inquiries about my background or experience, feel free to contact me.
Quick Overview
PhD researcher in safe and responsible AI, working on making AI systems interpretable, robust, and equitable. My work includes detecting and mitigating bias, stress-testing models under real-world distribution shift, and building interpretable-by-design architectures that are human-in-the-loop. Applied primarily to skin cancer detection, my methods have delivered state-of-the-art, fairness-improved, and clinically auditable models, published across journals and conferences including an early-accepted (top 9%) MICCAI 2026 paper. I am currently extending this work to cardiac vision-language models and validating clinically. I bring hands-on experience translating research into reproducible, well-documented tools and training materials for interdisciplinary researchers.
Education
-
PhD in Computer Science (2023-2026 Expected)
Keele University - Competition-funded Department Studentship
Thesis: Responsible AI for Healthcare: Robustness, Interpretability, and Fairness in Skin Lesion Recognition
Viva: Scheduled 3rd July 2026
2 first-author publications (PAA 2025, AIiH 2025 Oral), 1 first-author paper early accepted (Top 9%) at MICCAI 2026, 1 first-author paper under review (Image and Vision Computing), 2 co-authored accepted papers -
MSc in Artificial Intelligence and Data Science (2021-2022)
Keele University - Distinction
Data Analyst Internship, Turing Network Development Research Associate -
BSc in Physics (2018-2021)
Keele University - First Class Honours
Key Research Projects
- DermFormer: Transformer-based multi-modal architecture for robust skin cancer detection with state-of-the-art performance on Derm7pt dataset
- FairCBM: Fairness-aware curriculum learning for Concept Bottleneck Models, reducing demographic performance disparities by 44% and improving lowest-group outcomes by 63% (Early Accepted, MICCAI 2026)
- SCAN-CBM: Sparse pairwise polynomial classifier capturing concept absence patterns for interpretable diagnostic reasoning (Under Review, Image and Vision Computing)
- Zero-Shot Segmentation: CAM-guided foundation models (SAM/MedSAM) for automated dermoscopy analysis on ISIC datasets (35k+ images)
- SimDrift: Educational platform for visualising ML model drift across 24+ pre-trained models and 15+ realistic drift scenarios on MedMNIST
- Bias Mitigation: Cardiovascular disease mortality prediction fairness analysis using multi-center MINAP dataset (400k+ patients)
Publications
- 2 Published First-Author Papers: Pattern Analysis and Applications (2025), AIiH 2025 Conference (Oral Presentation)
- 1 First-Author Paper: Early Accepted (Top 9%) at MICCAI 2026 — Fair Curriculum Learning for CBMs in Dermatology
- 1 First-Author Paper: Under Review — Image and Vision Computing (SCAN-CBM)
- 2 Co-Authored Accepted Papers: Discover Computing, Springer Nature; AIiH 2026 (ClinAuditAI)
- 1 Co-Authored Paper Under Review: Machine Vision and Applications (Zero-Shot Crack Segmentation)
Technical Skills
- Programming: Python, LaTeX, R, SQL
- ML Frameworks: PyTorch, TensorFlow, Keras, scikit-learn, Hugging Face
- Frontier Models: Segment Anything Model (SAM/MedSAM), vision-language models (VLMs), zero-shot transfer evaluation
- Computer Vision: OpenCV, torchvision, TIMM
- Explainable AI: CAM, GradCAM, Integrated Gradients, attention visualisation, SHAP, concept bottleneck models
- Fairness & Bias: Aequitas, Fairlearn, demographic parity, equalized odds, adversarial debiasing, fairness benchmarking
- Medical Imaging: MONAI, SimpleITK, dermoscopy analysis, clinical validation
- Development: Git/GitHub, Jupyter, Weights & Biases, HPC cluster management, SLURM
- Certifications: BlueDot Technical AI Safety Course
Professional Service
- Programme Committee: International Conference on AI in Healthcare (AIiH) 2026
- Special Session Organiser: Co-organising session on Explainability & Accountability at AIiH 2026
- Peer Review: Engineering Applications of Artificial Intelligence (2024–Present); 1st International Workshop on AI Safety and Security (AI-SS) 2026
- Leadership: PGR Representative (3 years), Student Representative (3 years)
- Teaching: Module Lead — Level 7 NHS Apprenticeship, Image Processing (2026–Present); Laboratory Demonstrator (4 years); Guest Lecturer on Deep Learning & Responsible AI
For the most up-to-date version of my CV, please download the PDF above. For specific inquiries about my background or experience, feel free to contact me.