Searcharxiv⌕ Search
Home/Archive/arXiv

arXiv · 2610.11418

Experience-Guided Initiation Search for Learned Skills in Skill Composition

Qixuan Li·Yanhong Zhao·Jincheng Yu
View original source ↗Read source document ↗Export metadata ↓

Abstract

Deploying frozen learned skills, such as Vision-Language-Action (VLA) policies, in new environments requires identifying initiation configurations that support reliable execution. In skill composition, an initiation configuration affects not only the current skill but also the physical state passed to subsequent skills, so successful execution of an individual skill does not necessarily imply successful completion of the composed task. Estimating target-specific capability through extensive rollouts is costly in real-world deployment, while directly reusing historical experience can be unreliable under environment changes. We propose EVIS, an Experience-Guided and Behavior-Validated Initiation Search framework for discovering reliable initiation configurations under limited target interaction. EVIS uses historical execution experience to prioritize promising candidates and target-environment behavior to validate whether they remain effective. We evaluate EVIS on single-skill and two-stage manipulation tasks with frozen VLA policies. EVIS reduces mean target-environment queries and improves reliable candidate discovery under small interaction budgets. These results show that combining historical guidance with target-side behavioral validation can reduce the interaction cost of deploying frozen learned skills in new environments.

Explore related subjects

cs.RO

Keep this discovery

Explore connections, maps & timelines

BibTeXRIS

Qixuan Li, Yanhong Zhao, Jincheng Yu. 2026-10-08. Experience-Guided Initiation Search for Learned Skills in Skill Composition. https://arxiv.org/abs/2610.11418

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related papers

arXiv2026

The Kinetics Observer: A Tightly Coupled Estimator for Legged Robots

This paper presents the Kinetics Observer, a novel proprioceptive state estimator for legged robots designed to provide the feedback required for versatile locomotion and physical interaction with the environment, with enhanced robustness to contact slippage. Its core contribution is a tight coupling between whole-body kinematics and external wrenches through a contact dynamics model, enabling the consistent fusion of leg kinematics, IMU, and wrench sensor measurements for real-time joint estimation of centroidal kinematics, contact rest poses, contact wrenches, and disturbance wrenches. By exploiting contact wrench measurements as correction terms, the proposed observer increases sensing redundancy, which provides observability of contact slippage relative to the centroid frame and enhanced robustness to modeling and sensor errors. The approach is experimentally evaluated on two humanoid robots across three scenarios totaling thirteen walking sequences, including long-distance walking with repeated contact changes, locomotion over slippery obstacles, and non-coplanar multicontact motion.

cs.RO↗
arXiv2026

RoboAug: One Annotation to Hundreds of Scenes via Region-Contrastive Data Augmentation for Robotic Manipulation

Enhancing the generalization of robotic learning in diverse unseen environments remains a fundamental challenge. Existing approaches often rely on large-scale pretraining, which is labor-intensive and time-consuming, or semantic data augmentation methods that assume flawless upstream object detection in real-world scenarios. In this work, we propose RoboAug, a novel generative data augmentation framework that reduces reliance on large-scale pretraining and perfect visual recognition by requiring only a single image with bounding box annotations for dataset construction. Leveraging this minimal supervision, RoboAug employs pretrained generative models for precise semantic augmentation and introduces a plug-and-play region-contrastive loss to guide attention toward task-relevant regions, thereby enhancing generalization and task success rates. Extensive real-world experiments on UR-5e, AgileX, and Tian Gong 2.0 demonstrate that RoboAug consistently outperforms state-of-the-art augmentation baselines under background, distractor, and lighting shifts. Our project is available at https://x-roboaug.github.io/.

cs.RO↗
arXiv2026

ActionCodec: What Makes for Good Action Tokenizers

Vision-Language-Action (VLA) models leveraging the native autoregressive paradigm of Vision-Language Models (VLMs) have demonstrated superior instruction-following and training efficiency. Central to this paradigm is action tokenization, yet its design has primarily focused on reconstruction fidelity, failing to address its direct impact on VLA optimization. Consequently, the fundamental question of \textit{what makes for good action tokenizers} remains unanswered. In this paper, we bridge this gap by establishing design principles specifically from the perspective of VLA optimization. We identify a set of best practices based on information-theoretic insights, including maximized temporal token overlap, minimized vocabulary redundancy, enhanced multimodal mutual information, and token independence. Guided by these principles, we introduce \textbf{ActionCodec}, a high-performance action tokenizer that significantly enhances both training efficiency and VLA performance across diverse simulation and real-world benchmarks. Notably, on LIBERO, a SmolVLM2-2.2B fine-tuned with ActionCodec achieves a 95.5\% success rate without any robotics pre-training. With advanced architectural enhancements, this reaches 97.4\%, representing a new SOTA for VLA models without robotics pre-training. We believe our established design principles, alongside the released model, will provide a clear roadmap for the community to develop more effective action tokenizers.

cs.RO↗