Research map / Video / Video generation
MilliVid: Adaptive Latents for Long-Range Consistency in Video Generation
Ishaan Chandratreya, David Charatan, Basile Van Hoorick, Sergey Zakharov, Vitor Guizilini, Phillip Isola, Vincent Sitzmann
Read the paperCheck your idea against 137,217 papers
AI-extracted summary. Written by AI from the paper's abstract only. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Proposes MilliVid, which generates long videos via coarse-to-fine rollout in a multi-scale adaptive latent token space to improve long-horizon consistency under manageable sequence lengths.
Problem
Long-horizon consistency in transformer-based video generation is hard because even a few dozen frames produce impractically long token sequences.
Approach
Pre-trains an adaptive autoencoder that compresses video into multi-scale latent tokens and generates video by coarse-to-fine rollout within this multi-scale token space, mitigating sequence-length growth.
Closest papers by meaning
- WorldWeaver: Generating Long-Horizon Video Worlds via Rich Perception (NeurIPS 2025 · similarity 0.88)
- Snap Video: Scaled Spatiotemporal Transformers for Text-to-Video Synthesis (CVPR 2024 · similarity 0.87)
- AdaState: Self-Evolving Anchors for Streaming Video Generation (NeurIPS 2026 · similarity 0.87)
- StreamingT2V: Consistent, Dynamic, and Extendable Long Video Generation from Text (CVPR 2025 · similarity 0.87)
- Generative Inbetweening through Frame-wise Conditions-Driven Video Generation (CVPR 2025 · similarity 0.87)
- ARLON: Boosting Diffusion Transformers with Autoregressive Models for Long Video Generation (ICLR 2025 · similarity 0.87)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
