Join us for the Genμ Challenge on Generative Machine Unlearning and U&ME Workshop @ ICCV 2025
Mayank Vatsa
Genμ Challenge and Unlearning and Model Editing Workshop at ICCV 2025
Paper submission and Generative Machine Unlearning Challenge
We’re excited to share that we’re organizing the Genμ Challenge on Generative Machine Unlearning, as part of the U&Me Workshop at ICCV 2025, taking place in Honolulu, Hawaii (October 19–20).
This challenge is all about pushing the boundaries of what it means to forget in large generative models. Participants will explore how to remove specific concepts, like a particular object, identity, or style, from text-to-image diffusion models, while still preserving general knowledge and generation quality.
Here’s what you need to know:
The challenge is built on Stable Diffusion v1.4.
We’re releasing the Genμ benchmark, which includes Forget, Locality, and Adjacency sets, designed specifically for this task.
The prompt-level CSVs are already available (no images, to ensure ethical and legal compliance).
Submission deadline: August 1, 2025
Top teams will be invited to co-author a challenge summary paper.
If your work touches on unlearning, model editing, diffusion models, or responsible AI, we’d love to see you participate!
👉 https://unlearning.iab-rubric.org
You can also consider submitting your work to the U&Me Workshop @ ICCV 2025. Details here:
👉 https://sites.google.com/view/u-and-me-workshop/home
Feel free to share this with colleagues who might be interested. If you have any questions, don’t hesitate to reach out: genmu.chal...@gmail.com
We’re looking forward to seeing what the community creates!
Best regards,
Mayank Vatsa, PhD
Fellow IEEE, IAPR
Professor and Swarnajayanti Fellow
IIT Jodhpur, India
Adjunct Professor IIIT Delhi
Associate EiC, IEEE T-BIOM
Member, National Academy of AI (USA)
http://home.iitj.ac.in/~mvatsa/
http://iab-rubric.org/