Multimodal Agent MemorySubmit a paper

Context-as-Memory

Context as Memory: Scene-Consistent Interactive Long Video Generation with Memory Retrieval

Jiwen Yu, Jianhong Bai, Yiran Qin, Quande Liu, Xintao Wang, Pengfei Wan, et al. · 2025-08-12

Read the paperPDF

Management
T2 Dynamic Memory
Memory changes during deployment through a stable management policy.
Primary modality
Video & streaming
Also: text
Representation
Source- and modality-specific

How experience becomes memory

Each path is one memory component, traced from what is stored to how it is found again.

  1. historical generated frames

    1. Represented asRetained perceptual source evidenceSource- and modality-specific
    2. Stored inFlat sequential records
    3. Retrieved byMetadata / spatiotemporal retrieval
    4. Source evidenceSource directly available
    Historical frames are stored without post-processing and retrieved by field-of-view overlap.Full-paper method / memory representation and retrieval · PDF p. 2 · Figure 1

How memory changes (T2 operations)

  • Update · Rule-based update

    Each generated segment appends new frames to the historical context memory under a fixed rule. For Context-as-Memory, this specifically supports fixed-rule insertion or replacement of historical generated frames.

Related systems