PhD student, National Technical University of Athens
1 paper at NeurIPS 2025
Jointly modeling image laments and high-level semantic features (e.g. DINO), significantly enhances generative quality and training efficiency.
