Skip to content
Preprint

TaskAnchor: Grounding Task State in Reactive VLAs for Long-Horizon Manipulation

Sep 2026 · 0 citations
Computer Science

Abstract

Reactive vision--language--action (VLA) models struggle with long-horizon manipulation when visually similar observations can correspond to different actions depending on the task stage or interaction history. We refer to this ambiguity as task-state aliasing and introduce TaskAnchor, a lightweight adapter that grounds pretrained VLAs in execution history. TaskAnchor combines history-conditioned visual refinement with a milestone-supervised task-state coordinate, a scalar representing the semantic stage of execution. These signals are injected through the native visual and language interfaces, respectively, without introducing an explicit planner or modifying the action-generation mechanism. On RMBench, TaskAnchor achieves approximately 4.9--5.5$\times$ the average success rates of the published $\pi_{0.5}$ and X-VLA baselines, with consistent gains on RoboMemArena and real robots. The added latency is only 2.08\,ms per action chunk for $\pi_{0.5}$.

View source

Similar papers

Preprint Oct 2026

EvoMem-VLA: State-Evolution Memory for Long-Horizon Robot Manipulation

Most vision-language-action (VLA) models rely on current observations and lose task-relevant evidence once it leaves view, limiting performance on long-horizon, memory-dependent tasks. Existing efforts incorporate compressed historical features or sparse visual keyframes. However, isolated snapshots can leave the polic...

Yu-Heng Na, Zhi-De Zhong, Jun-Jie He et al. · 1 citation
Preprint Aug 2026

Decoding Task Progress from VLA Representations

The results suggest that VLAs have rich, linearly readable internal representations of semantic quantities like task progress, and that learning to read these signals offers a lightweight, interpretable path toward monitoring deployed visuomotor policies.

Atiksh Bhardwaj, E. W. Duan, Prithwish Dan et al. · 3 citations
Preprint Sep 2026

Causeway: Restoring Task Accessibility for Instruction Switching in VLA Policies

Vision-language-action (VLA) policies can execute many tasks from standard initial states, yet a new instruction may fail after another task has altered the robot's physical state. We study instruction switching, where a new task is issued during or after the execution of a different one. We observe that a target task...

Qing-Zi Wang, Kai Feng, Guang-Yao Shi et al. · 0 citations
Sep 2026

SEES: A Self-Evolving Embodied System via Failure-Guided VLA Policy Adaptation

A Self-Evolving Embodied System (SEES) that learns from failures and improves the VLA policy without additional expert demonstrations, and can be integrated with different VLA backbones to progressively improve their long-horizon performance.

Zi-Wen Li, Hanlue Zhang, Zhen-Yang Ren et al. · 1 citation
Preprint Aug 2026

TemporalFlow-VLA: Learning Physically Grounded Execution History for Long-Horizon Robot Manipulation

TemporalFlow-VLA provides a compact, physically grounded interface for exploiting ordered execution history without explicit motion estimation or geometric processing at deployment, and shows its clearest advantage over prior methods on longer-horizon, multi-stage manipulation.

Jia-Rui Yang, Ye-Hao Lu, Yu-Ning Su et al. · 2 citations

We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.