Machine learning is at the core of AI innovations in many domains, such as health, engineering, business, safety and security. It enables computers to realise problem-solving capabilities that frequently match or even surpass humans.
In the DMB group, we develop machine learning for Computer Vision, Biometrics, Data Engineering and other fields. Computer Vision aims at automated interpretation, analysis, and understanding of visual information from images and videos. Biometrics enables automated recognition of persons using traits like face, fingerprint, iris, and vein patterns. Data Engineering is about obtaining high quality data for development and use of AI through automated collection, integration, transformation, and cleaning of data.
Regulations and real-world innovations pose strong requirements on the technology in terms of transparency, explainability, robustness, trustworthiness, fairness, data quality, resilience to attacks, privacy, sustainability, and efficiency. Our research specifically focuses on these important aspects for developing responsible AI.
Examples are explainable AI-support for medical diagnosis and treatment decisions, face recognition resilient to face morphing attacks, understanding of soil ecology through network data science, counter energy-hungry training of LLMs and deep learning with sparse models, process mining for healthcare, understanding social interactions in first-person videos, identifying creators of artwork through biometric traces like fingerprints, etc.
DMB offers the specialisations Computer Vision and Biometrics, which is open to EE and EMSYS students and Data Science and Technology for CS students. Courses taught by DMB inlcude Advanced COmputer Vision and Pattern Recognition, Introduction to Biometrics, Data Science, and Machine Learning. Student's assigments (both MSc and BSc) are defined as real research projects, i.e. they actively contribute to DMB's research on e.g. detection of image manipulation and deep fakes, recognition of faces in painings, surgical video analysis, text mining from medical files, sparse neural networks, embedded AI and more. Good MSC and BSc reports are often published at conferences/workshops or even in scientific journals.
