Multimodal Agent MemorySubmit a paper

One Sentence One Drama

One Sentence, One Drama: Personalized Short-Form Drama Generation via Multi-Agent Systems

Yufei Shi, Weilong Yan, Naixuan Huang, Yucheng Chen, Chenyu Zhang, Tao He, et al. · 2026-05-21

Read the paperPDF

Management
T2 Dynamic Memory
Memory changes during deployment through a stable management policy.
Primary modality
Video & streaming
Also: text, Image, Embodied & 3D
Representation
Hybrid
Distinct operative representation paths are deliberately linked or jointly consumed by downstream reasoning, generation, or control.

How experience becomes memory

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

  1. narrative plan

    1. Represented asModality-specific encodings & geometric stateSource- and modality-specific
    2. Stored inHierarchical multi-tier store
    3. Retrieved bySimilarity search, Metadata / spatiotemporal retrieval, Structure traversal
    4. Source evidenceRetained but not accessible to the agent
    Multi-agent debate outputs; 3D-grounded first-frame reference; reviewer-loop error reports; transition and BGM planning records. retrieves/reuses the shared spatial reference and stage-level reviewer feedback during visual/video generation and targeted revision.Full-paper method / memory representation and retrieval · PDF p. 2 · Figure 1 / method overview

  2. shared 3D-grounded spatial reference

    1. Represented asPerception-to-text records & summariesCross-represented explicit
    2. Stored inHierarchical multi-tier store
    3. Retrieved bySimilarity search, Metadata / spatiotemporal retrieval, Structure traversal
    4. Source evidenceSource not retained
    Multi-agent debate outputs; 3D-grounded first-frame reference; reviewer-loop error reports; transition and BGM planning records. retrieves/reuses the shared spatial reference and stage-level reviewer feedback during visual/video generation and targeted revision.Full-paper method / memory representation and retrieval · PDF p. 2 · Figure 1 / method overview

  3. cross-clip character/scene state

    1. Represented asSpatial, temporal & cross-modal structured memoryCross-represented explicit
    2. Stored inHierarchical multi-tier store
    3. Retrieved bySimilarity search, Metadata / spatiotemporal retrieval, Structure traversal
    4. Source evidenceSource not retained
    Multi-agent debate outputs; 3D-grounded first-frame reference; reviewer-loop error reports; transition and BGM planning records. retrieves/reuses the shared spatial reference and stage-level reviewer feedback during visual/video generation and targeted revision.Full-paper method / memory representation and retrieval · PDF p. 2 · Figure 1 / method overview

How memory changes (T2 operations)

  • Update · Prompted LLM/VLM record update

    Revision loops update script/visual/video outputs; spatial reference maintains consistent layout across clips; no long-term cross-session memory. For One Sentence One Drama, this specifically supports LLM-mediated memory editing of narrative plan.

  • Revision · Evidence-weighted belief revision

    Revision loops update script/visual/video outputs; spatial reference maintains consistent layout across clips; no long-term cross-session memory. For One Sentence One Drama, this specifically supports evidence-weighted revision of stored belief or confidence of narrative plan.

Related systems