LILI Lab
We are the Learners of Interpretable Latent Information (LILI) Lab, based in the Department of Informatics, University of Sussex. We create interpretable ML models for ”inverse problems”, where the information of interest is not directly measurable.
We aim to build models that are:
- Useful, providing meaningful insights to our interdisciplinary collaborators.
- Aware of uncertainty and communicate this.
- Contributing trustworthy tools for healthcare applications
What We Do
Opportunities
Ways to work with us or collaborate:
- Working in our lab: expectations, skills, and resources for joining.
- Collaborating with us: how we scope projects with partners and keep work reproducible.
Continual Learning
Reading groups, book clubs, and teaching highlights.
Software
Research tooling and reproducible codebases.
Outreach
Engagement like science fairs, and international teaching projects.
Our Research Themes
Medical Image Analysis
Work on MRI reconstruction, image registration, and related inverse problems.
Progression Modelling
Probabilistic latent variable modelling of neurodevelopment and degeneration.
Computer Vision
Projects spanning visual recognition, reconstruction, and interpretable models for vision tasks.
Ecological Monitoring
Interpretable learning for ecological data, from species monitoring to environmental sensing.
Machine Learning Methods
Core methodological work on probabilistic modeling, interpretability, and efficient training for inverse problems.