Xindi Wu(@cindy_x_wu):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. 🔗 https://t.co/TbKXjQMN3H 1/10

Xindi Wu

@cindy_x_wu

PhD student @PrincetonCS | Data-centric multimodal ml | prev @roboVisionCMU @CMU_Robotics | @NVIDIAAI @RealityLabs @Snapchat

加入 May 2017

1.5K 正在關注 5.3K 粉絲

Xindi Wu@cindy_x_wu

2026.01.20 20:03

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