PhD Position: Generative & Dynamic AI for Disease Progression Modeling
Position Type: Full-time PhD Student
Field: Biomedical Engineering / Computer Science / Electrical & Computer Engineering / Applied AI
Location: University of Alberta / Alberta Machine Intelligence Institute
Focus: Longitudinal Deep Learning, Generative & Dynamic Modeling, Medical Imaging
About the Project
Modern clinical care relies heavily on static imaging snapshots, which frequently fail to capture the complex temporal dynamics, velocity, and acceleration of chronic disease progression. Standard deep learning models excel at single-timepoint classification, but real-world clinical data poses unique challenges: patients are monitored at irregular time intervals, imaging data contains inherent noise and artifacts, and individual patient trajectories vary widely.
This research initiative focuses on advancing the state-of-the-art in AI-driven disease progression modeling. By shifting the paradigm from static diagnostic tools to continuous spatio-temporal forecasting, the project aims to predict pathological trajectories long before irreversible clinical outcomes occur. The successful candidate will develop and apply advanced machine learning frameworks spanning standard deep learning architectures, generative diffusion and flow models, and continuous-time dynamical systems like Neural ODEs to model high-dimensional longitudinal medical data and map disease dynamics under real-world clinical conditions.
Key Responsibilities
- Design, implement, and optimize deep learning architectures (e.g., Neural ODEs, continuous-time diffusion or flow models, Time-Aware Transformers) capable of handling irregularly sampled time-series imaging data.
- Integrate Evidential Deep Learning (EDL) and Conformal Prediction layers to produce per-pixel uncertainty maps and confidence intervals, distinguishing between true tissue progression and imaging artifacts.
- Combine complex 3D volumetric latent representations with longitudinal clinical metrics to model disease conversion and progression.
- Scale and train complex models across GPU clusters, utilizing efficient gradient techniques (e.g., adjoint sensitivity methods) and optimized latent space operations.
- Publish research findings in top-tier machine learning and medical imaging venues (e.g., MICCAI, NeurIPS, IEEE TMI, CVPR) and present at international conferences.
Required Qualifications & Technical Expertise
- M.Sc. (or equivalent) in Computer Science, Biomedical Engineering, Electrical Engineering, Applied Mathematics, or a related quantitative field.
- Strong theoretical foundation and hands-on experience in training deep neural network architectures using PyTorch.
- Experience or strong background in diffusion models, flow-based models, continuous-time networks (Neural ODEs/SDEs), or temporal sequence modeling.
- Hands-on experience working with 3D volumetric medical images (e.g., OCT, MRI, CT) using standard tools (e.g., MONAI, ITK, SimpleITK, DICOM/NIfTI formats).
- Extensive experience in Linux/Unix environments, shell scripting, and distributed training on high-performance compute clusters (e.g., SLURM, multi-GPU scaling).
Desired Attributes & Personal Traits
- A drive to explore emerging AI paradigms, tackle complex computational and mathematical challenges (e.g., solver instabilities, non-Euclidean latent dynamics), and rapidly pick up new methodologies.
- Ability to diagnose model convergence issues, manage class imbalances, and implement statistical reliability measures.
- Strong written and verbal communication skills to work effectively within an interdisciplinary team bridging computer science, engineering, and clinical specialists.
What We Offer
- Access to dedicated high-performance GPU clusters and infrastructure.
- Opportunity to conduct research directly impacting patient care and addressing major healthcare challenges.
- Close collaboration with leading researchers in AI, medical imaging, and clinical specialists at the University of Alberta and at Amii.
How to Apply
Interested applicants should submit the following:
- Cover Letter (1–2 pages) highlighting your technical fit, research interests, and motivation for joining this project.
- Curriculum Vitae (CV) including a list of publications, prior projects, and technical skills.
- Portfolio Link (e.g., GitHub profile, link to a paper implementation or M.Sc. thesis repo).
Please fill out this form and email the cover letter and CV to Craig Jones (cjones7@ualberta.ca).
For more information about the lab, see the University of Alberta directory listing.