Coordination, research and development on ML analysis with subgroup discovery for trustworthy AI. Collaboration with CISPA on an ELSA (European Lighthouse on Secure and Safe AI) grant.

We perform research on making deep neural networks more interpretable and robust to adversarial input perturbations.

Harnessing deep networks to improve our understanding of the human vision system. We investigated the use of deep networks as analogy to the human vision system and showed that trained on two unrelated tasks, the network builds two separate information processing streams, akin to the human brain.

Detection, Segmentation and Tracking of Brain Metastasis in MRI scans using Deep Convolutional Networks.

Development of methods for video based detection:- detect and locate pedestrian crossings on streets- detect unmarked road boundaries and estimate driving lane via a maximum a posteriori method

R&D on video based detection of beacons on construction sites.

Assisting lecturers in a series of computer engineering lectures.