Liangchen Song

Researcher@Apple·美国·OpenReview

研究方向

diffusion models · 3D generation · NeRF · dynamic scenes · neural rendering · implicit representation · detection · segmentation · recognition · domain adaptation · knowledge distillation

6.5

CAR-Flow: Condition-Aware Reparameterization Aligns Source and Target for Better Flow Matching

NeurIPS 2025Spotlight

三作

5.8

Cavia: Camera-controllable Multi-view Video Diffusion with View-Integrated Attention

ICLR 2025Rejected

5.5

Cavia: Camera-controllable Multi-view Video Diffusion with View-Integrated Attention

合作者 (20)

Liangchen Song

AToken: A Unified Tokenizer for Vision

Score Distillation of Flow Matching Models

Autoregressive Video Generation beyond Next Frames Prediction

CAR-Flow: Condition-Aware Reparameterization Aligns Source and Target for Better Flow Matching

Cavia: Camera-controllable Multi-view Video Diffusion with View-Integrated Attention

Cavia: Camera-controllable Multi-view Video Diffusion with View-Integrated Attention