Seyedmorteza Sadat

Ph.D. student at DisneyResearch|Studios & ETH Zurich

profile.jpg

Hi there! I am a joint Ph.D. student between DisneyResearch|Studios and ETH Zurich supervised by Prof. Otmar Hilliges, Prof. Thomas Hofmann, and Dr. Romann Weber.

I specialize in diffusion models for image and video generation. My research focuses on developing novel sampling and guidance techniques to improve the quality, diversity, and efficiency of diffusion models. More broadly, I am interested in how diffusion models can be trained and tuned for visual content creation, as well as their role in creative applications and world modeling.

Prior to my Ph.D., I obtained my master’s degree in Computer Science from ETH Zurich and my bachelor’s degree in Electrical Engineering from Sharif University of Technology.

news

Dec 05, 2022 I started my Ph.D. at DisneyResarch|Studios and ETH Zurich.

selected publications

  1. ICLR 2026
    HiGS: History-Guided Sampling for Plug-and-Play Enhancement of Diffusion Models
    Seyedmorteza Sadat, Farnood Salehi, and Romann M. Weber
    In The Fourteenth International Conference on Learning Representations, 2026
  2. arXiv
    Guidance in the Frequency Domain Enables High-Fidelity Sampling at Low CFG Scales
    Seyedmorteza Sadat, Tobias Vontobel, Farnood Salehi, and Romann M. Weber
    2025
  3. TMLR
    Efficient Distillation of Classifier-Free Guidance using Adapters
    Cristian Perez Jensen and Seyedmorteza Sadat
    Transactions on Machine Learning Research, 2025
  4. NeurIPS 2025
    Token Perturbation Guidance for Diffusion Models
    Javad Rajabi, Soroush Mehraban, Seyedmorteza Sadat, and Babak Taati
    In The Thirty-ninth Annual Conference on Neural Information Processing Systems, 2025
  5. SIGGRAPH Asia 2025
    HiWave: Training-Free High-Resolution Image Generation via Wavelet-Based Diffusion Sampling
    Tobias Vontobel, Seyedmorteza Sadat, Farnood Salehi, and Romann M. Weber
    2025
  6. ICLR 2025
    Eliminating Oversaturation and Artifacts of High Guidance Scales in Diffusion Models
    Seyedmorteza Sadat, Otmar Hilliges, and Romann M. Weber
    In The Thirteenth International Conference on Learning Representations, 2025
  7. ICLR 2025
    No Training, No Problem: Rethinking Classifier-Free Guidance for Diffusion Models
    Seyedmorteza Sadat, Manuel Kansy, Otmar Hilliges, and Romann M. Weber
    In The Thirteenth International Conference on Learning Representations, 2025
  8. NeurIPS 2024
    LiteVAE: Lightweight and Efficient Variational Autoencoders for Latent Diffusion Models
    Seyedmorteza Sadat, Jakob Buhmann, Derek Bradley, Otmar Hilliges, and Romann M. Weber
    In The Thirty-eighth Annual Conference on Neural Information Processing Systems, 2024
  9. ICLR 2024
    CADS: Unleashing the Diversity of Diffusion Models through Condition-Annealed Sampling
    Seyedmorteza Sadat, Jakob Buhmann, Derek Bradley, Otmar Hilliges, and Romann M. Weber
    In The Twelfth International Conference on Learning Representations, 2024