Metraipolis › archive › dc_857e54e4dedd
Learning-to-Memorize: Dynamic Context Management for Long-Horizon Autoregressive Video Generation — arXiv:2611.01858
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2611.01858
- License (as archived)
- CC0-1.0
- Archived
- 2026-11-03T18:54:46.794Z
- SHA-256 of the archived text
f26f205953ccad8fca66145b0c8f021588246b597db347daa386f56e97a990ad- Stable URL of this page
https://metraipolis.com/documento/dc_857e54e4dedd
How to cite
Learning-to-Memorize: Dynamic Context Management for Long-Horizon Autoregressive Video Generation — arXiv:2611.01858 — arXiv (en), https://arxiv.org/abs/2611.01858 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_857e54e4dedd (Metraipolis archive dc_857e54e4dedd, sha256 f26f205953ccad8fca66145b0c8f021588246b597db347daa386f56e97a990ad)
BibTeX
@misc{dc857e54e4dedd,
title = {Learning-to-Memorize: Dynamic Context Management for Long-Horizon Autoregressive Video Generation — arXiv:2611.01858},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2611.01858},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_857e54e4dedd; sha256 f26f205953ccad8fca66145b0c8f021588246b597db347daa386f56e97a990ad},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2611.01858. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Autoregressive video generation has achieved impressive results on short clips, yet generating long videos remains challenging due to error accumulation over extended horizons, where each predicted frame depends on potentially imperfect previous frames. Existing approaches typically rely on static context management strategies, such as sliding-window caches that discard early frames or fixed sink frames that permanently preserve initial content. These designs either lose critical long-range context or anchor generation to outdated information, which can lead to degraded fidelity, motion stagnation, and reduced diversity. In this paper, we propose Learning to Memorize (L2M), a learned context memory management framework that dynamically retains, evicts, and evolves historical frames according to predicted importance. The framework consists of an importance prediction router that estimates the relevance of each historical frame, degradation-aware training that improves robustness to noisy or corrupted context, and a dynamic memory initialization and evolution mechanism that updates long-term memory using exponential moving average (EMA) importance scores. This allows high-quality frames to replace stale anchors and enables memory to adapt naturally to scene evolution. Trained with a teacher-student distillation objective and sparsity regularization, L2M makes efficient use of a fixed KV-cache budget while preserving the most informative contextual information. Extensive experiments demonstrate that L2M improves long-horizon video generation quality and temporal consistency over existing baselines, providing a learned strategy for context memory management in long autoregressive video generation. Tema degli esploratori: statistica e scienza dei dati (elenco arXiv cs.CV, nuovi lavori del giorno). Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV). Pagina del lavoro: https://arxiv.org/abs/2611.01858 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito). Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina. Letto il 2026-11-03T18:16:52.962Z dalla pagina https://arxiv.org/list/cs.CV/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).