Multimodal Agent MemorySubmit a paper

LaMem-VLA

Dual Latent Memory in Vision-Language-Action Models for Robotic Manipulation

Hongyu Qu, Jianzhe Gao, Xiaobin Hu, Shaohuan Yang, Xinlei Yu, Rui Yan, et al. · 2026-07-08

Read the paperPDF

Management
T2 Dynamic Memory
Memory changes during deployment through a stable management policy.
Primary modality
Embodied & 3D
Also: Image, 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. short- and long-term latent vaults

    1. Represented asPersistent multimodal latent stateInternalized latent & parametric
    2. Stored inHierarchical multi-tier store, Vector-indexed collection
    3. Retrieved bySimilarity search, Latent / parametric activation
    4. Source evidenceSource not retained
    Curator, seeker, condenser, and weaver manage latent experience for VLA reasoning.Full-paper method / memory representation and retrieval · PDF p. 5 · Figure 1

How memory changes (T2 operations)

  • Update · Learned update module

    During operation, LaMem-VLA applies a trained but frozen update module to short- and long-term latent vaults; the memory state changes while the governing update policy remains fixed.

  • Consolidation · Merging & deduplication

    The condenser compacts retrieved evidence. For LaMem-VLA, this specifically supports merging or deduplication of overlapping entries of short- and long-term latent vaults.

  • Forgetting · Discrete eviction

    During operation, LaMem-VLA applies explicit deletion or eviction of stored items to short- and long-term latent vaults; the memory state changes while the governing update policy remains fixed.

Related systems