arXiv Machine Learning
When Plans Change, Remember: Replanning-Guided Visual Memory for Vision-Language-Action Models
Read the original on arXiv Machine Learning →The Flow has not summarised this story yet — read it at arXiv Machine Learning.
Related stories
arXiv AI
5d ago
Memory Without Curation: Learning Multi-Stream Latent Memory for Robot Policies
arXiv:2610.14406v1 Announce Type: cross Abstract: It is widely appreciated that robotic policies require long-term memory to perform complex tasks. However, directly conditioning policies on all past...
arXiv AI
Oct 2
MIKASA-Robo-VLA: Benchmarking Memory in VLA Models for Long-Horizon Manipulation
arXiv:2610.00604v1 Announce Type: cross Abstract: Vision-language-action policies often see only one or a few recent frames, which makes it difficult to evaluate how they use information that disappe...
arXiv Computer Vision
Oct 1
D$^2$-VLA: Dual-Memory Dual-Frequency Vision-Language-Action Model For Long Dynamic Manipulation
arXiv:2609.34792v2 Announce Type: replace Abstract: Long-horizon manipulation requires robots to remember cues that are no longer in view while responding to moving objects. Yet vision-language-actio...
arXiv AI
Oct 7
Adapting Vision-Language-Action Models to Unknown Visual Disruptions During Execution
arXiv:2610.07946v1 Announce Type: cross Abstract: Visual disruptions can arise while a robot is executing a task, leaving a vision-language-action (VLA) policy to respond without knowing the disrupti...
arXiv Computer Vision
Sep 30
EventVLA: Event-Driven Visual Evidence Memory for Long-Horizon Vision-Language-Action Policies
arXiv:2606.20092v3 Announce Type: replace Abstract: Memory remains a critical bottleneck for long-horizon robotic manipulation, as standard Vision-Language-Action (VLA) policies often fail when task-...
arXiv AI
Oct 2
Divide-and-Remember: Recursive Action-Relevant Memory for Long-Horizon VLA Policies
The paper introduces Divide-and-Remember (D&R), a recursive memory method for vision-language-action (VLA) policies that optimises memory by maximizing the conditional mutual information between actions and memory given observations. D&R recursively divides the full history into top‑K selections over 2K tokens, using a shared lightweight selector across all recursion blocks to handle unbounded histories efficiently. Evaluated on the RoboMME benchmark of 16 long‑horizon manipulation tasks, D&R achieves state‑of‑the‑art success rates with consistent gains across all suites while using only 64 tokens, and similar improvements are observed in real‑robot experiments.