Stephan Günnemann

A new method streamlines pruning of neural networks for improved performance.

New methods improve language model safety while keeping functionality intact.

A new method improves communication in graph data processing.

New method reveals vulnerabilities in GNN explanation methods.

New GNN architecture improves predictions through enhanced expressiveness and fragment integration.

Neural wave functions and Pfaffians enhance quantum chemistry predictions significantly.

A new framework improves molecular graph generation techniques for drug discovery.

This article discusses safeguarding GNNs from data poisoning and backdoor attacks.

A study on the resilience of Graph Transformers against adversarial attacks.

Integrating quantum computing and randomized smoothing improves machine learning model security against attacks.

This article discusses improving the robustness of quantum machine learning models against adversarial attacks.

A look at activation steering for improving AI memory handling.