影响力指数

45.54/100

前 10.6%

全站排名 #6,848

发表论文6 篇

平均评分5.4

年均产出3.0 篇/年

Shanchuan Lin

Researcher@ByteDance Inc.·美国·OpenReview

研究方向

diffusion models · computer vision · matting

SkipSR: Faster Super-Resolution with Token Skipping

ICLR 2026Desk Rejected

SeedVR2: One-Step Video Restoration via Diffusion Adversarial Post-Training

ICLR 2026Poster

VINCIE: Unlocking In-context Image Editing from Video

ICLR 2026Poster

Growing Visual Generative Capacity for Pre-Trained MLLMs

ICLR 2026Withdrawn

Autoregressive Adversarial Post-Training for Real-Time Interactive Video Generation

NeurIPS 2025Poster

Diffusion Adversarial Post-Training for One-Step Video Generation

ICML 2025Poster

合作者 (20)