Time-to-Move: Training-Free Motion Controlled Video Generation via Dual-Clock Denoising
ICLR 2026 · ICLR ReALM-GEN Workshop · CVPR J2A Workshop
Assaf Singer
PhD student in Computer Science at the Technion, advised by Dr. Or Litany. I work on generative AI, 3D/4D reconstruction, video generation, and tracking.
About me
I am a PhD student at the Technion - Israel Institute of Technology, advised by Dr. Or Litany. My research focuses on controllable video generation and dynamic 3D/4D scene understanding - how to reconstruct, interact with, and generate dynamic scenes for next-generation autonomous systems.
Previously, I was a computer vision researcher at the Weizmann Institute of Science with Prof. Tali Dekel, where I worked on point tracking. I hold an M.Sc. in Computer Science from NYU, advised by Prof. Kyunghyun Cho and Prof. Katharina Kann, and a B.Sc. in Mathematics from Tel Aviv University. Before academia, I spent three years as a data scientist at Pagaya Technologies, building ML models for credit risk and loan evaluation.
News
- Oral talk at Israel Computer Vision Day on Time-to-Move.
- Time-to-Move accepted to ICLR 2026, and to the ICLR ReALM‑GEN and CVPR J2A workshops!
- Oral talk at Israel Computer Vision Day on DINO-Tracker.
- Started PhD at the Technion, advised by Dr. Or Litany.
- DINO-Tracker accepted to ECCV 2024!
Publications
DINO-Tracker: Taming DINO for Self-Supervised Point Tracking in a Single Video
ECCV 2024
Earlier work
Talks
Oral presentation - Time-to-Move: Training-Free Motion Controlled Video Generation via Dual-Clock Denoising.
Oral presentation - DINO-Tracker: Taming DINO for Self-Supervised Point Tracking in a Single Video.
Contact
For inquiries or collaboration, please email me at assaf.singer@campus.technion.ac.il.