Xindi Wu
@cindy_x_wu
PhD student @PrincetonCS | Data-centric multimodal ml | prev @roboVisionCMU @CMU_Robotics | @NVIDIAAI @RealityLabs @Snapchat
1.5K Following 5.3K Followers
New #NVIDIA# Paper
We introduce Motive, a motion-centric, gradient-based data attribution method that traces which training videos help or hurt video generation.
By isolating temporal dynamics from static appearance, Motive identifies which training videos shape motion in video generation.
🔗
1/10
Show more
