UniMPA: A Unified Memory-Prediction-Action Model via Action-Grounded Transition Modeling

UniMPA: A Unified Memory-Prediction-Action Model via Action-Grounded Transition Modeling

|Authors: Wei Li, Rui Shao, Jie He, Lingsen Zhang, Ziwei Liu, Liqiang Nie

|Venue: arXiv preprint | Year: 2026

|arXiv: 2609.11875v1

Abstract

Recent advances in Vision-Language-Action (VLA) models have improved robotic manipulation, yet observation-to-action learning remains limited by a fundamental transition realizability gap, manifested in three tightly coupled problems: (i) Transition ambiguity. Visually similar current observations may correspond to different manipulation phases and imply different subsequent transitions. (ii) Prediction—execution mismatch. A visually plausible predicted future observation does not necessarily correspond to a physically realizable transition. (iii) Experience—realization mismatch. A historically executable action pattern may not necessarily realize the intended transition in the current scene and therefore requires context-aware adaptation. Accordingly, we propose UniMPA, a Unified Memory-Prediction-Action model that addresses these problems through a shared action-grounded transition interface. (i) UniMPA introduces Persistent-Selective Future Prediction to resolve transition ambiguity by modeling the intended future state evolution. A persistent latent stream continuously tracks task-level progress, while a transition-critical pixel stream selectively resolves fine-grained interaction changes through memory-grounded prediction. (ii) To assess the physical executability of the anticipated transition, the predicted transition queries a temporal Visual-Action Memory Bank. The bank retrieves historically realized visual-action experience, grounding future prediction in executable evidence. (iii) To adapt executable experience to the current scene, an Action-Visual Memory Bank retrieves visually grounded action prototypes from historical action evolution. Prototype-Biased Flow then shifts the flow source toward a historically supported action manifold for context-aware refinement.

Key Contributions

• Recent advances in Vision-Language-Action (VLA) models have improved robotic manipulation, yet ob…
• Visually similar current observations may correspond to different manipulation phases and imply d…
• (ii) Prediction—execution mismatch
• A visually plausible predicted future observation does not necessarily correspond to a physically…
• (iii) Experience—realization mismatch

Topics

• manipulation
• vision
• vla
• reinforcement-learning

Code & Data

• GitHub repository: https://github.com/JiuTian-VL/UniMPA

BibTeX

@article{Li2026_260911875v1,
  title     = {UniMPA: A Unified Memory-Prediction-Action Model via Action-Grounded Transition Modeling},
  author    = {Wei Li and Rui Shao and Jie He and Lingsen Zhang and Ziwei Liu and Liqiang Nie},
  year      = {2026},
  eprint    = {2609.11875v1},
  archivePrefix = {arXiv},
  primaryClass  = {cs.RO},
  url       = {https://arxiv.org/abs/2609.11875v1}
}