معرفی
Jannik Wolff is a Researcher and Ph.D. Candidate at TU Berlin's Machine Learning Group, advised by Prof. Klaus-Robert Müller and Shinichi Nakajima. He is also a Research Associate at the Berlin Institute for the Foundations of Learning and Data (BIFOLD). Previously, he worked at SAP AI Research. His research focuses on generative models for multimodal data, probabilistic modeling, and applications in computer vision and inverse problems. He holds a Master’s degree from TU Berlin and a patent for graph-based priors in zero-shot learning.
Key research interests include generative models (e.g., diffusion models, transformers), probabilistic ML, and multimodal learning. He has published peer-reviewed work on hierarchical VAEs, mixture-of-experts models, and low-shot learning techniques.
Notable contributions include a patent on graph-based priors for image classification and a drone delivery system optimized via hybrid behavior planning. His work on Bayesian deep learning for optimization and emotion recognition using EEG data further highlights his interdisciplinary expertise.
