Guardado en:
Detalles Bibliográficos
Autores principales: Mahdizadeh, Ailar, Azadi, Puria, Li, Muchen, He, Xiangteng, Sigal, Leonid
Formato: Preprint
Publicado: 2026
Materias:
Acceso en línea:https://arxiv.org/abs/2605.14310
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866917494284877824
author Mahdizadeh, Ailar
Azadi, Puria
Li, Muchen
He, Xiangteng
Sigal, Leonid
author_facet Mahdizadeh, Ailar
Azadi, Puria
Li, Muchen
He, Xiangteng
Sigal, Leonid
contents Streaming video understanding with large vision-language models (VLMs) requires a compact memory that can support future reasoning over an ever-growing visual history. A common solution is to compress the key-value (KV) cache, but existing streaming methods typically rely on local token-wise heuristics, such as recency, temporal redundancy, or saliency, which do not explicitly optimize whether the retained cache is representative of the accumulated history. We propose to view KV-cache compression as a coreset selection problem: rather than scoring tokens independently for retention, we select a small subset that covers the geometry of the accumulated visual cache. Our method operates in a joint KV representation and introduces a bicriteria objective that balances coverage in key and value spaces, preserving both retrieval structure and output-relevant information. To encourage a more diverse retained subset, we further introduce an orthogonality-driven diversity criterion that favors candidates contributing new directions beyond the current selection, and connect this criterion to log-determinant subset selection. Across four open-source VLMs and five long-video and streaming-video benchmarks, our method improves over heuristic streaming compression baselines under a fixed cache budget. These results highlight that representative coreset selection offers a more effective principle, than token-wise pruning, for memory-constrained streaming video understanding.
format Preprint
id arxiv_https___arxiv_org_abs_2605_14310
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle CoRDS: Coreset-based Representative and Diverse Selection for Streaming Video Understanding
Mahdizadeh, Ailar
Azadi, Puria
Li, Muchen
He, Xiangteng
Sigal, Leonid
Computer Vision and Pattern Recognition
Streaming video understanding with large vision-language models (VLMs) requires a compact memory that can support future reasoning over an ever-growing visual history. A common solution is to compress the key-value (KV) cache, but existing streaming methods typically rely on local token-wise heuristics, such as recency, temporal redundancy, or saliency, which do not explicitly optimize whether the retained cache is representative of the accumulated history. We propose to view KV-cache compression as a coreset selection problem: rather than scoring tokens independently for retention, we select a small subset that covers the geometry of the accumulated visual cache. Our method operates in a joint KV representation and introduces a bicriteria objective that balances coverage in key and value spaces, preserving both retrieval structure and output-relevant information. To encourage a more diverse retained subset, we further introduce an orthogonality-driven diversity criterion that favors candidates contributing new directions beyond the current selection, and connect this criterion to log-determinant subset selection. Across four open-source VLMs and five long-video and streaming-video benchmarks, our method improves over heuristic streaming compression baselines under a fixed cache budget. These results highlight that representative coreset selection offers a more effective principle, than token-wise pruning, for memory-constrained streaming video understanding.
title CoRDS: Coreset-based Representative and Diverse Selection for Streaming Video Understanding
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2605.14310