视频大模型

Vista: Scene-Aware Optimization for Streaming Video Question Answering under Post-Hoc Queries

Vista：面向事后查询的场景感知流视频问答优化

Haocheng Lu, Nan Zhang, Wei Tao, Xiaoyang Qu, Guokuan Li, Jiguang Wan, Jianzong Wang

专题命中 视频问答 ：video understanding(abstract)；分类 cs.CV

AI总结 Vista通过场景感知分割、压缩和召回技术，实现了高效且可扩展的流视频问答，提升了长上下文推理能力，同时保持低延迟和内存效率。

Comments Accepted to AAAI 2026 (Main Technical Track)