LVMem: Image-Augmented Memory for Long-Horizon Question Answering

Key points

• Unlike conventional visual question answering, realistic robotic settings require reasoning over extended temporal horizons and persistent scene memory.1
• We build upon the view-based memory paradigm and propose an efficient framework for long-horizon spatio-temporal question answering.1
• 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.1
• 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.1

Sources (1)

• [1]LVMem: Image-Augmented Memory for Long-Horizon Question Answering
  arXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 19, 06:01 PM

  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.

Extractive summary: sentences quoted from the sources.