4 papers
Diffusion Models Memorize in Training -- and Generalize in Inference
Tim Kaiser, Markus Kollmann
Diffusion models generalize well in practice. However, an optimal diffusion model fully memorizes the training data and therefore fails to generalize, raising the question of what…
ClusterMine: Robust Label-Free Visual Out-Of-Distribution Detection via Concept Mining from Text Corpora
Nikolas Adaloglou, Diana Petrusheva, Mohamed Asker +2
Large-scale visual out-of-distribution (OOD) detection has witnessed remarkable progress by leveraging vision-language models such as CLIP. However, a significant limitation of cur…
Guiding a diffusion model using sliding windows
Nikolas Adaloglou, Tim Kaiser, Damir Iagudin +1
Guidance is a widely used technique for diffusion models to enhance sample quality. Technically, guidance is realised by using an auxiliary model that generalises more broadly than…
Rethinking cluster-conditioned diffusion models for label-free image synthesis
Nikolas Adaloglou, Tim Kaiser, Felix Michels +1
Diffusion-based image generation models can enhance image quality when conditioned on ground truth labels. Here, we conduct a comprehensive experimental study on image-level condit…