Research map / Video / Video generation
Learning-to-Memorize: Dynamic Context Management for Long-Horizon Autoregressive Video Generation
Read the paperCheck your idea against 108,967 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 a learned dynamic context management approach for long-horizon autoregressive video generation to mitigate error accumulation from static cache strategies.
Problem
Long autoregressive video generation suffers from error accumulation because each frame depends on imperfect previous frames; existing static context strategies (sliding-window caches, fixed sink frames) are limited.
Approach
Replaces static context management rules with a learned policy that dynamically decides which past frames to keep in memory during autoregressive generation, as suggested by the title and abstract (details not available).
Experimental setup
Baselines: sliding-window cache fixed sink frames
Closest papers by meaning
- Attend Locally, Remember Linearly: Linear Attention as Cross-Frame Memory for Autoregressive Video Diffusion (NeurIPS 2026 · similarity 0.89)
- StreamingT2V: Consistent, Dynamic, and Extendable Long Video Generation from Text (CVPR 2025 · similarity 0.89)
- Addressable Memory for Video World Models (NeurIPS 2026 · similarity 0.89)
- OmniMem: Scalable and Adaptive Memory Retrieval for Long Video Generation (NeurIPS 2026 · similarity 0.89)
- Self Forcing: Bridging the Train-Test Gap in Autoregressive Video Diffusion (NeurIPS 2025 · similarity 0.88)
- VideoMAR: Autoregressive Video Generation with Continuous Tokens (NeurIPS 2025 · similarity 0.88)
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.