Multimodal data modeling, Efficient inference methods, Deep uncertainty estimation
Dr. Shinichi Nakajima leads the Independent Research Group Probabilistic Modeling and Inference. His aim is to develop novel probabilistic models and inference methods for multimodal, heterogeneous, and complex structured data analysis. In particular he wants to provide machine learning tools that can incorporate multiple aspects of data samples observed under different circumstances, in efficient and theoretically grounded ways. This includes:
- developing novel probabilistic models with efficient inference methods
- exploring novel applications of probabilistic models, and
- establishing uncertainty estimation methods for deep probabilistic models.
Thomas Schnake, Oliver Eberle, Jonas Lederer, Shinichi Nakajima, Kristof T. Schütt, Klaus-Robert Müller, Gregoire Montavon
Higher-Order Explanations of Graph Neural Networks via Relevant Walks
Lorenz Vaitl, Kim Andrea Nicoli, Shinichi Nakajima, Pan Kessel
Gradients should stay on Path: Better Estimators of the Reverse- and Forward KL divergence for Normalizing Flows
Lorenz Vaitl, Kim Andrea Nicoli, Shinichi Nakajima, Pan Kessel
Path-Gradient Estimators for Continuous Normalizing Flows
Kirill Bykov, Anna Hedström, Shinichi Nakajima, Marina M.-C. Höhne
NoiseGrad: enhancing explanations by introducing stochasticity to model weights
Ping Xiong, Thomas Schnake, Gregoire Montavon, Klaus-Robert Müller, Shinichi Nakajima
Efficient Computation of Higher-Order Subgraph Attribution via Message Passing

Research Group Lead

Doctoral Researcher