Multimodal Agent MemorySubmit a paper

MAViS

MAViS: A Multi-Agent Framework for Long-Sequence Video Storytelling

Qian Wang, Ziqi Huang, Ruoxi Jia, Paul Debevec, Ning Yu · 2026-01-24

Read the paperPDF

Management
T2 Dynamic Memory
Memory changes during deployment through a stable management policy.
Primary modality
Video & streaming
Also: text
Representation
Internalized latent & parametric

How experience becomes memory

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

  1. Per-character identity (LoRA via MUDI)

    1. Represented asExperience-conditioned parametric memoryInternalized latent & parametric
    2. Stored inModular adapter bank
    3. Retrieved byPlanner / tool-assisted retrieval, Latent / parametric activation
    4. Source evidenceSource not retained
    Per-character LoRA weights. Character LoRA is applied per shot to keep protagonists consistent.Full paper: Implementation Details · PDF p. 14 · Figure 1 / method overview

How memory changes (T2 operations)

  • Update · Optimization-based update

    Per-character LoRA is learned (a fixed adaptation mechanism). For MAViS, this specifically supports a fixed operational optimization objective of Per-character identity (LoRA via MUDI).

Related systems