aiqbench analysis
This research introduces SARI, a novel framework that optimizes training for Vision-Language-Action models by splitting tasks into simulated free-space approaches and real-world contact interactions. By leveraging synthetic data for spatial diversity while reserving physical demonstrations for critical contact physics, the method significantly reduces data collection costs and improves generalization to unseen object placements. This advancement addresses a key bottleneck in deploying robust robotic manipulation systems that require precise handling of physical objects.
SARI: Phase-Split Sim-Real Co-Training for Contact-Rich Manipulation
arXiv:2610.02804v1 Announce Type: new Abstract: Vision-language-action (VLA) models often require costly real-world demonstrations to adapt to contact-rich manipulation tasks, particularly when generalization across object placements is needed. We propose SARI (Simulated Approach, Real Interaction), a phase-split sim-and-real co-training framework built on a simple insight: spatial coverage and contact physics should be acquired from the domains best suited to them. Specifically, free-space approaches require spatial diversity but tolerate modest simulation gaps, making them ideal for synthetic generation; conversely, contact interactions demand accurate physics but vary little across object placements, allowing a few real demonstrations to generalize across the workspace. SARI generates diverse simulated approaches in a photorealistic digital twin while collecting real contact interactions at only a few placements. Post-trained on these phase-segmented demonstrations, a single policy seamlessly stitches simulated approaches with real contact interactions using visual appearance alignment and a shared camera-relative action representation--without explicit phase labels or hand-coded switches. Across five real-world contact-rich manipulation tasks, SARI reduces real-data collection time by 34.3% and achieves 27.5% success at unseen placements, where all full-task sim-real baselines fail completely (0%).
Abstract:Vision-language-action (VLA) models often require costly real-world demonstrations to adapt to contact-rich manipulation tasks, particularly when generalization across object placements is needed. We propose SARI (Simulated Approach, Real Interaction), a phase-split sim-and-real co-training framework built on a simple insight: spatial coverage and contact physics should be acquired from the domains best suited to them. Specifically, free-space approaches require spatial diversity but tolerate modest simulation gaps, making them ideal for synthetic generation; conversely, contact interactions demand accurate physics but vary little across object placements, allowing a few real demonstrations to generalize across the workspace. SARI generates diverse simulated approaches in a photorealistic digital twin while collecting real contact interactions at only a few placements. Post-trained on these phase-segmented demonstrations, a single policy seamlessly stitches simulated approaches with real contact interactions using visual appearance alignment and a shared camera-relative action representation--without explicit phase labels or hand-coded switches. Across five real-world contact-rich manipulation tasks, SARI reduces real-data collection time by 34.3% and achieves 27.5% success at unseen placements, where all full-task sim-real baselines fail completely (0%).
Robotics (cs.RO)
arXiv:2610.02804 [cs.RO]
(or arXiv:2610.02804v1 [cs.RO] for this version)
https://doi.org/10.48550/arXiv.2610.02804 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 266 words
Related coverage
Building a Foundation Stack for General-Purpose Robots
IEEE Spectrum · Jul 13, 2026
Game Design Creativity evals
LessWrong · Oct 12, 2026
ForeTime-VLA: Causal Future-Token Distillation from a World Action Model for Conveyor-Belt Manipulation
arXiv · Aug 24, 2026
Task aware Dynamic Movement Primitives for failure detection and recovery in contact rich manipulation
arXiv · Sep 22, 2026