Miran Ozdogan
Miran Ozdogan
Interests
I'm a first year DPhil student in Computer Science, working in the OxMedIS research group.
I previously completed an integrated Bachelors in Computer Science and Statistics at the Ludwig-Maximilians-University of Munich, where I researched the automated optimization of a tree-based reduction method for multi-class classification called nested dichotomy as my final major project. Subsequently, I completed my MPhil in Cambridge where I developed a new tree-based ensembling method that can optimize arbitrary loss functions.
My research interests are:
- AutoML for ensembling
- Multi-class classification
- Pretraining methods for medical visual representations
Selected Publications
-
Cooperative Co−Evolution for Ensembles of Nested Dichotomies for Multi−Class Classification
Marcel Wever‚ Miran Özdogan and Eyke Hüllermeier
2023.
Details about Cooperative Co−Evolution for Ensembles of Nested Dichotomies for Multi−Class Classification | BibTeX data for Cooperative Co−Evolution for Ensembles of Nested Dichotomies for Multi−Class Classification