摘要：对于具身 Agent 而言，可靠的语言引导感知至关重要，它们必须根据长期的观测历史来回答有关物体、事件及其位置的问题。与传统的视觉问答（VQA）不同，现实的机器人场景需要跨越更长的时间跨度进行推理并具备持久的场景记忆能力。我们在基于视角的记忆（view-based memory）范式基础上，提出了一种用于长时程时空问答的高效框架。现有的基于视角的系统通常仅存储文本描述（captions）或视觉嵌入（visual embeddings），而我们扩展了类似 ReMEmbR 的记忆机制，加入了图像路径，将每个存储的描述与其原始视觉内容关联起来。为了使存储的注释更具信息量，我们在记忆构建过程中用基于光流的帧选择策略（PSFR）替代了均匀采样，并存储以物体为中心的描述性场景表征，而非简单的整幅图像标注。此外，我们还为基于 VLM 的 Agent 配备了自适应检索工具，可返回与最相关描述关联的图像，从而允许 Agent 自行决定何时需要进行视觉检查。在 OC-NaVQA 基准测试中，LVMem 在两种骨干网络下的二元指标和位置指标上均优于 ReMEmbR，并在二元准确率上追平或超越了 4D 场景图基线 DAAAM，同时它仅依赖轻量级的基于视角的记忆运行，无需构建显式图结构。

Abstract:Reliable language-guided perception is essential for embodied agents, where questions about objects, events, and their locations must be answered from long observation histories. Unlike conventional visual question answering, realistic robotic settings require reasoning over extended temporal horizons and persistent scene memory. We build upon the view-based memory paradigm and propose an efficient framework for long-horizon spatio-temporal question answering. While existing view-based systems typically store either text captions or visual embeddings, we extend the ReMEmbR-style memory with image paths that link each stored description to its original visual content. To make the stored annotations more informative, we replace uniform sampling during memory construction with an optical-flow-based frame selection strategy (PSFR) and store an object-centric descriptive scene representation instead of plain captions. We further equip a VLM-based agent with an adaptive retrieval tool that returns the image associated with the most relevant caption, allowing the agent to decide when visual inspection is needed. On the OC-NaVQA benchmark, LVMem outperforms ReMEmbR on binary and positional metrics with both backbones and matches or exceeds a 4D scene graph DAAAM baseline on binary accuracy, while operating on a lightweight view-based memory that requires no explicit graph construction.