Multimodal Agent MemorySubmit a paper

DreamRunner

DreamRunner: Fine-Grained Compositional Story-to-Video Generation with Retrieval-Augmented Motion Adaptation

Zun Wang, Jialu Li, Han Lin, Jaehong Yoon, Mohit Bansal · 2025-11-14

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. Retrieval-augmented motion priors

    1. Represented asExperience-conditioned parametric memoryInternalized latent & parametric
    2. Stored inModular adapter bank, Vector-indexed collection
    3. Retrieved bySimilarity search, Metadata / spatiotemporal retrieval, Latent / parametric activation
    4. Source evidenceSource not retained
    Retrieved video motion references; Test-time LoRA weights. BM25 then CLIP retrieval over the InternVid corpus selects motion references.Full paper: Diffusion Model · PDF p. 4 · Figure 1 / method overview

How memory changes (T2 operations)

  • Update · Optimization-based update

    Per-scene/character LoRA is adapted at test time (a fixed adaptation mechanism). For DreamRunner, this specifically supports a fixed operational optimization objective of Retrieval-augmented motion priors.

Related systems