KeyRec: Bounded Visual Memory for Streaming and Long-Video Understanding

Zihan chen, Xuejian Rong, Xiaojuan Wang, Boqing Gong, Adi Zicher, Yael Pritch, Nikhil Karnad

KeyRec: Bounded Visual Memory for Streaming and Long-Video Understanding: 2 upvotes on Hugging Face Daily Papers, #55 of 58 papers on 2026-10-05. Day-by-day upvote history.

Vision-language models are increasingly used to understand long videos and continuous streams. However, dense visual tokens accumulate with video duration, making long-context inference prohibitively expensive. Existing training-free visual-token selection methods reduce this cost by retaining informative tokens, but may lose coherent event evidence and fail to distinguish detailed recent observations from long-range history. We propose KeyRec, a training-free framework for constructing bounded visual memory. During query-agnostic writing, KeyRec preserves fine-grained recent observations in a visual cache and organizes historical evidence into a structured event bank. Candidate events are proposed according to their novelty relative to previously stored events and maintained through an online add--merge--evict update. When a question arrives, a text-only router adaptively allocates a fixed readout budget between recent and event memory, without reprocessing historical frames. KeyRec operates on model-facing visual embeddings and supports both modular encoder--projector VLMs and the encoder- and projector-free NEO-ov architecture. Across four streaming and long-video benchmarks and three VLM backbones, KeyRec achieves the best compressed performance in 13 of 15 settings using only 10\% of the dense decoder-facing visual-token budget. It outperforms the strongest compressed baseline by 2.21--18.37 points on real-time questions, achieves the best compressed result in five of six long-video settings, and performs best in every NEO-ov 2B setting.

Paper page on Hugging Face · arXiv

Data: hysts-bot-data/daily-papers-stats and the Daily Papers API. Open data: tardellirs/paper-pulse-data. Sister project: Model Pulse, the download history of every model on the Hub.