Full Professor, University of Copenhagen
2 papers at NeurIPS 2025
A novel framework for responsible text-to-image generation that incorporates a dual-module transformation on the intermediate bottleneck representations of diffusion models.
