Cover Image

Going Down Memory Lane: Scaling Tokens for Video Stream Understanding with Dynamic KV-Cache Memory

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Agarwal, Vatsal, Suri, Saksham, Gwilliam, Matthew, Kumar, Pulkit, Shrivastava, Abhinav
Format:	Preprint
Published:	2026
Subjects:	Computer Vision and Pattern Recognition
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Similar Items

Evolutionary Caching to Accelerate Your Off-the-Shelf Diffusion Model
by: Aggarwal, Anirud, et al.
Published: (2025)

Towards Multimodal Understanding via Stable Diffusion as a Task-Aware Feature Extractor
by: Agarwal, Vatsal, et al.
Published: (2025)

StreamMem: Query-Agnostic KV Cache Memory for Streaming Video Understanding
by: Yang, Yanlai, et al.
Published: (2025)

LARP: Tokenizing Videos with a Learned Autoregressive Generative Prior
by: Wang, Hanyu, et al.
Published: (2024)

UPLiFT: Efficient Pixel-Dense Feature Upsampling with Local Attenders
by: Walmer, Matthew, et al.
Published: (2026)

LiFT: A Surprisingly Simple Lightweight Feature Transform for Dense ViT Descriptors
by: Suri, Saksham, et al.
Published: (2024)

TeCoNeRV: Leveraging Temporal Coherence for Compressible Neural Representations for Videos
by: Padmanabhan, Namitha, et al.
Published: (2026)

HERMES: KV Cache as Hierarchical Memory for Efficient Streaming Video Understanding
by: Zhang, Haowei, et al.
Published: (2026)

Explaining the Implicit Neural Canvas: Connecting Pixels to Neurons by Tracing their Contributions
by: Padmanabhan, Namitha, et al.
Published: (2024)

Decouple and Cache: KV Cache Construction for Streaming Video Understanding
by: Pang, Zhanzhong, et al.
Published: (2026)

Utilization of Neighbor Information for Image Classification with Different Levels of Supervision
by: Jayatilaka, Gihan, et al.
Published: (2025)

Towards Understanding Best Practices for Quantization of Vision-Language Models
by: Das, Gautom, et al.
Published: (2026)

Trokens: Semantic-Aware Relational Trajectory Tokens for Few-Shot Action Recognition
by: Kumar, Pulkit, et al.
Published: (2025)

How to Design and Train Your Implicit Neural Representation for Video Compression
by: Gwilliam, Matthew, et al.
Published: (2025)

UVIS: Unsupervised Video Instance Segmentation
by: Huang, Shuaiyi, et al.
Published: (2024)

Do text-free diffusion models learn discriminative visual representations?
by: Mukhopadhyay, Soumik, et al.
Published: (2023)

Latent-INR: A Flexible Framework for Implicit Representations of Videos with Discriminative Semantics
by: Maiya, Shishira R, et al.
Published: (2024)

SlotMemory: Object-Centric KV Memory for Streaming Long-Video Generation
by: Dou, Weijia, et al.
Published: (2026)

Trajectory-aligned Space-time Tokens for Few-shot Action Recognition
by: Kumar, Pulkit, et al.
Published: (2024)

CacheFlow: Compressive Streaming Memory for Efficient Long-Form Video Understanding
by: Patel, Shrenik, et al.
Published: (2025)

Characterizing Motion Encoding in Video Diffusion Timesteps
by: Baherwani, Vatsal, et al.
Published: (2025)

NeRF-Aug: Data Augmentation for Robotics with Neural Radiance Fields
by: Zhu, Eric, et al.
Published: (2024)

A Video is Worth 10,000 Words: Training and Benchmarking with Diverse Captions for Better Long Video Retrieval
by: Gwilliam, Matthew, et al.
Published: (2023)

Accelerate High-Quality Diffusion Models with Inner Loop Feedback
by: Gwilliam, Matthew, et al.
Published: (2025)

StreamKV: Streaming Video Question-Answering with Segment-based KV Cache Retrieval and Compression
by: Chen, Yilong, et al.
Published: (2025)

Efficient Continuous Video Flow Model for Video Prediction
by: Shrivastava, Gaurav, et al.
Published: (2024)

Streaming Video Question-Answering with In-context Video KV-Cache Retrieval
by: Di, Shangzhe, et al.
Published: (2025)

XStreamVGGT: Extremely Memory-Efficient Streaming Vision Geometry Grounded Transformer with KV Cache Compression
by: Su, Zunhai, et al.
Published: (2026)

XStreamVGGT: Extremely Memory-Efficient Streaming Vision Geometry Grounded Transformer with KV Cache Compression
by: Su, Zunhai, et al.
Published: (2026)

LiveVLM: Efficient Online Video Understanding via Streaming-Oriented KV Cache and Retrieval
by: Ning, Zhenyu, et al.
Published: (2025)

MA-LMM: Memory-Augmented Large Multimodal Model for Long-Term Video Understanding
by: He, Bo, et al.
Published: (2024)

Continuous Video Process: Modeling Videos as Continuous Multi-Dimensional Processes for Video Prediction
by: Shrivastava, Gaurav, et al.
Published: (2024)

LEIA: Latent View-invariant Embeddings for Implicit 3D Articulation
by: Swaminathan, Archana, et al.
Published: (2024)

Memory-efficient Streaming VideoLLMs for Real-time Procedural Video Understanding
by: Chatterjee, Dibyadip, et al.
Published: (2025)

StreamForest: Efficient Online Video Understanding with Persistent Event Memory
by: Zeng, Xiangyu, et al.
Published: (2025)

Memory Helps, but Confabulation Misleads: Understanding Streaming Events in Videos with MLLMs
by: Zhang, Gengyuan, et al.
Published: (2025)

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

MARC: Memory-Augmented RL Token Compression for Efficient Video Understanding
by: Wu, Peiran, et al.
Published: (2025)

StreamOV: Streaming Omni-Video Understanding via Evidence-Guided Memory and Response Triggering
by: Xie, Ming, et al.
Published: (2026)

StreamMeCo: Long-Term Agent Memory Compression for Efficient Streaming Video Understanding
by: Wang, Junxi, et al.
Published: (2026)