4LC: 4D Latent Cache for Efficient Conditioning in Dynamic World Modeling

Key points

• For video generation in dynamic world modeling, observed static and dynamic content should be efficiently reused to provide conditions for repeated generation at different times and viewpoints.1
• As a composite cache, 4LC binds visual representations from multiple visual encoders (e.g., VAE, DINO, V-JEPA, VGGT) to shared geometry and motion, allowing the composite representations to describe the dynamic world with rich information.1
• Building on 4LC, we develop 4LC-Video to address multiple video generation tasks through different configurations of source observations, spatiotemporal requests, and historical reuse, including camera control, temporal control, novel-view synthesis, scene revisits, and long-term generation.1
• Our experiments show more efficient condition acquisition, effective spatiotemporal controllability, and scene consistency across tasks using a single 4LC-Video model without task-specific fine-tuning.1

Sources (2)

• [1]4LC: 4D Latent Cache for Efficient Conditioning in Dynamic World Modeling
  arXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 13, 02:47 AM

  We introduce 4LC, a 4D latent cache with composite visual representations and a flexible spatiotemporal access interface.

  For video generation in dynamic world modeling, observed static and dynamic content should be efficiently reused to provide conditions for repeated generation at different times and viewpoints.

• [2]4LC: 4D Latent Cache for Efficient Conditioning in Dynamic World Modeling
  Hugging Face Daily Papers · Oct 13, 12:00 AM · same content

Extractive summary: sentences quoted from the sources.