让AI像人一样长期记忆并检索视频内容，提升理解能力。
Visual Agentic Memory: Enabling Online Long Video Understanding via Online Indexing, Hierarchical Memory, and Agentic Retrieval
- 动态筛选关键视觉信息，分层存储并支持后续查询
- 在OVO-Bench上达68.41分，优于直接使用大模型的基线
- 适合需要长期视频分析的场景，如智能监控、纪录片理解
长视频理解不仅依赖大上下文窗口，更需能决定保留哪些视觉证据、长期可检索，并基于可追溯观察进行推理的记忆机制。我们提出无需训练的视觉代理记忆（VAM），包含三个组件：在线索引在流式输入下实现选择性证据保留；分层记忆将时间上下文与空间观测并行组织；代理检索在生成答案前搜索、检查并验证候选证据。在OVO-Bench上，VAM平均得分68.41，高于相同底层模型Gemini 3 Flash的67.46。在月尺度的MM-Lifelong数据集（105.6小时，51天）上，达到17.11%，仅次于GPT-5的ReMA（17.62%）。结果表明，将视觉记忆视为显式、可检视、可查询的基底，有助于长时程视频理解。代码已开源。
原文摘要 · Abstract (English)
Long video understanding requires more than large context windows. It also needs a memory mechanism that decides what visual evidence to retain, keeps it searchable over long horizons, and grounds later reasoning in recoverable observations rather than compressed latent state alone. We propose Visual Agentic Memory (VAM), a training-free framework with three components. Online Indexing supports selective evidence retention under streaming constraints. Hierarchical Memory organises retained evidence in a Parallel Representation that aligns temporal context with spatial observations. Agentic Retrieval searches, inspects, and verifies candidate evidence before producing a grounded answer. On OVO-Bench, VAM achieves the highest RT+BT average (68.41) across all reported baselines, improving over end-to-end use of the same underlying MLLM (Gemini 3 Flash, 67.46). On the month-scale split of MM-Lifelong train@month (105.6 hours over 51 days), VAM reaches 17.11%, second only to ReMA with GPT-5 (17.62%). These results suggest that long-horizon video understanding benefits from treating visual memory as an explicit, inspectable, and queryable substrate. Code is available at https://github.com/yiliu-li/Visual-Agentic-Memory.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。