Papers

Trending research and the full catalog - each paper linked to the benchmarks, methods, and models it introduces.

Filtered by domain: video-classificationClear

RIVER: A Real-Time Interaction Benchmark for Video LLMs

Yi Wang, LiMin Wang, Xiangyu Zeng et al. · 4 Mar 2026

Real-time video interaction benchmark evaluates online video comprehension capabilities through retrospective memory, live-perception, and proactive anticipation tasks, revealing limitations of offline models in real-time processing and long-term memory.

Image Understanding Language Modeling Video classification

PyVision-RL: Forging Open Agentic Vision Models via RL

Kaipeng Zhang, Chen Wei, Ming Li et al. · 24 Feb 2026

Reinforcement learning for agentic multimodal models often suffers from interaction collapse, where models learn to reduce tool usage and multi-turn reasoning, limiting the benefits of agentic behavior.

Reinforcement Learning Video classification