Research
Research Directions
 |
Deep Learning Techniques : We take a multi-disciplinary perspective to develop inductive biases to increase generalization capabilities of deep neural networks. Particularly, we invent methods which induce and exploit structure of the data, the model, and its training. |
 |
Explainable AI : We develop novel techniques for explaining how self-learning AI systems (like Deep Neural Networks) perform their tasks. We particularly focus on explanations that are understandable for non-experts while still being faithful. We use established methodology and models from neuroscience and adapt them to build and understand AI systems. |
| |
Music : We design models to learn and represent harmonic structures in music, combining deep learning with music theory. Our work enhances both the analytical and creative capabilities of neural networks across various music information retrieval tasks. |
 |
Speech : Our research integrates deep learning, signal processing, and human perception to advance inclusive speech technologies. We develop novel approaches for speech synthesis, stutter detection, and speech anonymization , with an emphasis on preserving naturalness, expressivity, and inclusivity. |
 |
Medical Imaging : We rethink medical image processing algorithms to improve diagnosis of patients. We focus especially on the refinement of X-ray image styles and their quality. |
 |
: We design BCI systems to support people in their daily activities and enhance their quality of life. We aim to design more efficient systems by leveraging AI at every step of BCI systems: Transforming brain activities into a meaningful representation, classifying and turning them into actions. |
|
|
Education : We research on how technology can enhance the learning experience and rethink the way we educate our students about AI. |
