Fabio Pizzati
I am a postdoctoral researcher at MBZUAI, working with Ivan Laptev. Previously, I was a postdoc at the Torr Vision Group at the University of Oxford, working with Phil Torr. I am also a member of ELLIS.
I did my PhD at Inria Astra-vision team, in Paris (France), where I was supervised by Raoul de Charette.
I am currently interested in large-scale video diffusion models. In particular, I care about video generation for robotics and embodied AI applications, extending to world modeling and physics-aware learning. I am also interested in AI safety and multi-agent LLM systems.
news
| Jun 10, 2025 | New preprint on arxiv in collaboration with BBC on image tampering localization. |
|---|---|
| May 1, 2025 | I will visit INRIA Paris for one month in June. |
| Apr 1, 2025 | DiTFlow is accepted at CVPR 25! |
| Dec 20, 2024 | MALT is a new preprint on arxiv about LLM multi-agent post-training. |
| Dec 1, 2024 | 3 new preprints on arxiv. |
| Oct 10, 2024 | I joined MBZUAI as a postdoc, keeping deep ties with Oxford. |
| Sep 20, 2024 | One paper accepted at EMNLP Findings on LLM merging and safety implications. |
| Jul 29, 2024 | New preprint on arxiv on zero-shot image editing. |
| Jul 1, 2024 | Two papers accepted at ECCV. One on safety of large-scale diffusion models, one analysis paper on self-supervised learning. |
| Jun 3, 2024 | Our position paper on risks and benefits of open source GenAI is accepted at ICML 2024 position track as oral. |
| May 2, 2024 | I am now a member of ELLIS. |
| Apr 15, 2024 | I am visiting KAUST for one month, hosted by Bernard Ghanem. |
| Mar 20, 2024 | New paper out on the importance of pretraining data in self-supervised learning. Arxiv. |
| Feb 29, 2024 | Material Palette has been accepted to CVPR 2024. |
| Feb 7, 2024 | SynthCLIP has been featured in the TLDR newsletter with over 1 million subscribers. |
| Feb 2, 2024 | We trained a CLIP model entirely on synthetic data. Here the paper on arXiv. |
| Nov 28, 2023 | New paper on arXiv on material extraction. Link. |
| Sep 8, 2023 | We organized an amazing workshop on Weakly Supervised Computer Vision at Deep Learning Indaba. Check it out. |
| Mar 2, 2023 | New paper accepted at T-PAMI! Check it out |
| Mar 1, 2023 | I am starting as a postdoc in Oxford in the TVG. |
| Nov 29, 2022 | I successfully defended my PhD thesis :) |
| Oct 31, 2022 | I gave a talk at the VASC seminar of CMU on physics-informed image translation. |
| Jul 3, 2022 | ManiFest has been accepted at ECCV 2022. Code is out! |
| May 18, 2022 | I was selected for CVPR 2022 doctoral consortium. |
| Mar 9, 2022 | New paper out in collaboration with Jean-François Lalonde! |
| Feb 4, 2022 | Our local domains work has been accepted to VISAPP and awarded the best paper award. |
| Jul 1, 2021 | I recently gave two invited talks at EPFL and GTTI workshop on physics-informed image translation. The EPFL talk is available online! |
| May 6, 2021 | I was invited for a talk at Valeo.ai on CoMoGAN. |
| Apr 28, 2021 | CoMoGAN code released!. |
| Mar 12, 2021 | CoMoGAN has been accepted as CVPR oral. Arxiv is out, code will be released soon. |
| Jul 3, 2020 | Paper accepted at ECCV 2020! |
| Jun 9, 2020 | I presented Model-based disentanglement at a talk at GdR ISIS. Slides are out. |
| Oct 23, 2019 | New arxiv preprint for the Domain Bridge paper. Link |
| Oct 12, 2019 | Paper accepted at WACV 2020! |
| Oct 1, 2019 | I joined the RITS team in Paris as a PhD student. |