Cover Image

Task-Aware KV Compression For Cost-Effective Long Video Understanding

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Qin, Minghao, Shu, Yan, Zhang, Peitian, Lun, Kun, Yuan, Huaying, Zhou, Juenjie, Xiao, Shitao, Zhao, Bo, Liu, Zheng
Format:	Preprint
Published:	2025
Subjects:	Computer Vision and Pattern Recognition Artificial Intelligence
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Similar Items

Video-XL-2: Towards Very Long-Video Understanding Through Task-Aware KV Sparsification
by: Qin, Minghao, et al.
Published: (2025)

Video-XL: Extra-Long Vision Language Model for Hour-Scale Video Understanding
by: Shu, Yan, et al.
Published: (2024)

Memory-enhanced Retrieval Augmentation for Long Video Understanding
by: Yuan, Huaying, et al.
Published: (2025)

Long Context Compression with Activation Beacon
by: Zhang, Peitian, et al.
Published: (2024)

MLVU: Benchmarking Multi-task Long Video Understanding
by: Zhou, Junjie, et al.
Published: (2024)

VideoExplorer: Think With Videos For Agentic Long-Video Understanding
by: Yuan, Huaying, et al.
Published: (2025)

Compressing Lengthy Context With UltraGist
by: Zhang, Peitian, et al.
Published: (2024)

A Multi-Task Embedder For Retrieval Augmented LLMs
by: Zhang, Peitian, et al.
Published: (2023)

Extensible Embedding: A Flexible Multipler For LLM's Context Length
by: Shao, Ninglu, et al.
Published: (2024)

Flexibly Scaling Large Language Models Contexts Through Extensible Tokenization
by: Shao, Ninglu, et al.
Published: (2024)

Video-XL-Pro: Reconstructive Token Compression for Extremely Long Video Understanding
by: Liu, Xiangrui, et al.
Published: (2025)

TimeScope: Towards Task-Oriented Temporal Grounding In Long Videos
by: Liu, Xiangrui, et al.
Published: (2025)

DynamicKV: Task-Aware Adaptive KV Cache Compression for Long Context LLMs
by: Zhou, Xiabin, et al.
Published: (2024)

M3-Embedding: Multi-Linguality, Multi-Functionality, Multi-Granularity Text Embeddings Through Self-Knowledge Distillation
by: Chen, Jianlv, et al.
Published: (2024)

MomentSeeker: A Task-Oriented Benchmark For Long-Video Moment Retrieval
by: Yuan, Huaying, et al.
Published: (2025)

Large Language Models as Foundations for Next-Gen Dense Retrieval: A Comprehensive Empirical Assessment
by: Luo, Kun, et al.
Published: (2024)

Are Long-LLMs A Necessity For Long-Context Tasks?
by: Qian, Hongjin, et al.
Published: (2024)

MuKV: Multi-Grained KV Cache Compression for Long Streaming Video Question-Answering
by: Xiao, Junbin, et al.
Published: (2026)

C-Pack: Packed Resources For General Chinese Embeddings
by: Xiao, Shitao, et al.
Published: (2023)

Does RAG Really Perform Bad For Long-Context Processing?
by: Luo, Kun, et al.
Published: (2025)

BGE Landmark Embedding: A Chunking-Free Embedding Method For Retrieval Augmented Long-Context Large Language Models
by: Luo, Kun, et al.
Published: (2024)

Progressive Video Condensation with MLLM Agent for Long-form Video Understanding
by: Yin, Yufei, et al.
Published: (2026)

Extending Llama-3's Context Ten-Fold Overnight
by: Zhang, Peitian, et al.
Published: (2024)

Task-Aware Encoder Control for Deep Video Compression
by: Ge, Xingtong, et al.
Published: (2024)

InfoFlow KV: Information-Flow-Aware KV Recomputation for Long Context
by: Teng, Xin, et al.
Published: (2026)

Forcing-KV: Hybrid KV Cache Compression for Efficient Autoregressive Video Diffusion Models
by: Ji, Yicheng, et al.
Published: (2026)

NestedKV: Nested Memory Routing for Long-Context KV Cache Compression
by: Chen, Hong, et al.
Published: (2026)

Efficient Long-Horizon GUI Agents via Training-Free KV Cache Compression
by: Zhou, Bowen, et al.
Published: (2026)

Effectively Compress KV Heads for LLM
by: Yu, Hao, et al.
Published: (2024)

PackKV: Reducing KV Cache Memory Footprint through LLM-Aware Lossy Compression
by: Jiang, Bo, et al.
Published: (2025)

Beyond RAG: Task-Aware KV Cache Compression for Comprehensive Knowledge Reasoning
by: Corallo, Giulio, et al.
Published: (2025)

ChunkKV: Semantic-Preserving KV Cache Compression for Efficient Long-Context LLM Inference
by: Liu, Xiang, et al.
Published: (2025)

OmniGen: Unified Image Generation
by: Xiao, Shitao, et al.
Published: (2024)

FiLA-Video: Spatio-Temporal Compression for Fine-Grained Long Video Understanding
by: Guo, Yanan, et al.
Published: (2025)

VISTA: Visualized Text Embedding For Universal Multi-Modal Retrieval
by: Zhou, Junjie, et al.
Published: (2024)

LongVU: Spatiotemporal Adaptive Compression for Long Video-Language Understanding
by: Shen, Xiaoqian, et al.
Published: (2024)

Pyramid Forcing: Head-Aware Pyramid KV Cache Policy for High-Quality Long Video Generation
by: Chen, Jiayu, et al.
Published: (2026)

KVSculpt: KV Cache Compression as Distillation
by: Jiang, Bo, et al.
Published: (2026)

LVC: A Lightweight Compression Framework for Enhancing VLMs in Long Video Understanding
by: Wang, Ziyi, et al.
Published: (2025)

Head-Aware KV Cache Compression for Efficient Visual Autoregressive Modeling
by: Qin, Ziran, et al.
Published: (2025)