Nadav Cohen

PhD Candidate in Reichman University

[Nadav Cohen @nadavc220]

#NeurIPS Reviewer/AC discussion phase. No reviewer made the effort to even comment and there are less than 36 hours left.

What a shame.
Conferences should at least make reviewer post-rebuttal discussions public to authors as it seems some papers might be rejected by negligence

[Nadav Cohen @nadavc220]

Project Page: https://t.co/n4pjLzHy1l

[Nadav Cohen @nadavc220]

Excited to share that our latest paper "Colorful-Noise" was accepted to #SIGGRAPH2026.
We study white Gaussian noise and show that although structureless, it can be guided to broad colors and structure, with zero training or overhead.
Visit our project page for more details -->

[Nadav Cohen @nadavc220]

People are afraid that "machines are about to take their jobs". Seems to me that they might take their friends first...

Networking with robots at #CVPR2025

[Nadav Cohen @nadavc220]

@DanielCohenOr1 Well deserved indeed 👏😺

[Daniel Cohen-Or @DanielCohenOr1]

So happy to hear that my friend and colleague Ariel Shamir has received the Eurographics Outstanding Technical Contributions Award! Well deserved 👏🎉

[Itamar Weiss @itamarwe]

היום OpenAI עדכנו את מודל יצירת התמונות של ChatGPT, וזאת הזדמנות מצוינת לשרשור סרטים ישראלים בסגנון אנימה.
מוזמנים לנחש את שם הסרט בתגובות.

[Nadav Cohen @nadavc220]

@nicolayr_ @oronnir11 Thanks Nicolay =]

[Nadav Cohen @nadavc220]

[1/5] I'm excited to share that our paper, "Conditional Balance: Improving Multi-Conditioning Trade-Offs in Image Generation," has been accepted to #CVPR2025! 😸🎸🎓

A huge thank you to my collaborators, @oronnir11 and Prof. Ariel Shamir, for their invaluable contributions.

[Nadav Cohen @nadavc220]

@linoy_tsaban @oronnir11 Thanks Linoy 🙂

[Nadav Cohen @nadavc220]

@Animation "Conditional Balance" has been excepted to #CVPR2025 🎉🎊🪅

[Nadav Cohen @nadavc220]

[5/5] This allows us to direct style and content conditionals only to the areas where they matter most, ultimately improving the conditioning results while reducing the amount of conditioning required—without any additional training!

[Nadav Cohen @nadavc220]

[4/5] By analyzing the existing trade-offs between these two aspects, we propose a method that investigates the sensitivity of the model's generation process to conditional inputs.