Metraipolis › archive › dc_35ffd11ac361
Memory as Weights: Internalizing Long-Term History for Streaming Videos — arXiv:2610.22070
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2610.22070
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-16T09:30:56.059Z
- SHA-256 of the archived text
b4ac65dd3e00516081b4d55c6f5af2a35b43fa85b20d004b63260b7d87948e78- Stable URL of this page
https://metraipolis.com/documento/dc_35ffd11ac361
How to cite
Memory as Weights: Internalizing Long-Term History for Streaming Videos — arXiv:2610.22070 — arXiv (en), https://arxiv.org/abs/2610.22070 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_35ffd11ac361 (Metraipolis archive dc_35ffd11ac361, sha256 b4ac65dd3e00516081b4d55c6f5af2a35b43fa85b20d004b63260b7d87948e78)
BibTeX
@misc{dc35ffd11ac361,
title = {Memory as Weights: Internalizing Long-Term History for Streaming Videos — arXiv:2610.22070},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2610.22070},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_35ffd11ac361; sha256 b4ac65dd3e00516081b4d55c6f5af2a35b43fa85b20d004b63260b7d87948e78},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.22070. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Streaming vision-language models (VLMs) must continuously integrate information from the past while remaining responsive to current observations and user queries. Existing memory mechanisms typically compress, retrieve, or cache historical observations and reintroduce them into the input context, forcing historical memory and current observations to compete for limited context. We propose MaLoW, a streaming memory framework that instead internalizes historical context into evolving lightweight parameter updates. MaLoW uses a recurrent HyperNetwork to incrementally consolidate incoming video observations into fixed-size LoRA weights, allowing historical information to be carried forward without revisiting the full video stream. In doing so, MaLoW represents history as weights, while keeping the current observation as input. To support training, we further construct MaLoW-Instruct, a large-scale streaming video dataset containing 381K question-answer pairs over 56K videos. We conduct extensive experiments to demonstrate that MaLoW achieves strong performance on OVO-Bench, StreamingBench, and StreamArena. On the backward split of OVO-Bench, MaLoW outperforms prior memory approaches by 3.9% to 11.8% across different video backbones while incurring only modest latency overhead. Beyond streaming video understanding, MaLoW also generalizes to vision-language-action models, improving the action success rate on Simpler WidowX by 9.3%. Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno). Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV) ; Artificial Intelligence (cs.AI); Machine Learning (cs.LG). Pagina del lavoro: https://arxiv.org/abs/2610.22070 — 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 2026-10-16T08:44:18.024Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).