AXIS: A Growable Community-Driven Data Engine for Scalable Robot Manipulation

2026-07-23

AI Summary

This work addresses the limitations of existing robotic manipulation learning data pipelines, which are often constrained by specialized hardware, centralized teleoperation, or fixed task sets, resulting in poor scalability and limited diversity. To overcome these challenges, the authors propose AXIS—a scalable, community-driven data engine that collects large-scale demonstrations via browser-based teleoperation and integrates automated task generation, success detection, trajectory smoothing, and multimodal augmentation to construct high-quality training data. AXIS introduces a novel community collaboration framework enabling automatic task expansion, data curation, augmentation, and systematic evaluation. The resulting dataset encompasses 207 tasks and over 50,000 trajectories. Continuous pretraining on AXIS improves the success rate of π₀.₅ by 5.8% and outperforms the RoboCasa365 model by 37.3%, demonstrating substantial generalization capabilities in perturbed environments.

Abstract

Learning effective robot manipulation policies requires diverse, high-quality demonstrations, yet existing data pipelines are often difficult to scale because they rely on specialized hardware, centralized operators, or fixed task suites. We present AXIS, a growable community-driven data engine and benchmark for scalable robot learning, which enables browser-based teleoperation for large-scale demonstration collection, automatically generates and validates new manipulation tasks, and transforms community-collected demonstrations into training-ready data through automated success checking, quality filtering, trajectory smoothing, and visual and physics-based augmentation. The AXIS dataset currently contains 207 diverse tasks and 50K+ trajectories. Meanwhile, AXIS organizes data into task snapshots and evaluates policies with a systematic held-out protocol. We compare vision-language-action (VLA) policies under a unified AXIS evaluation suite and analyze scaling behavior across different data volumes. Continual pretraining on AXIS substantially improves the overall success rate of $π_{0.5}$ by 5.8%, outperforms the model pretrained on RoboCasa365 by 37.3%, and exhibits consistent scaling with increasing data volume, with the largest gains observed under layout, sensor-noise, and camera perturbations.

Problem

• scalable robot manipulation
• data pipeline
• demonstration collection
• task diversity
• robot learning

Innovation

• community-driven data engine
• scalable robot manipulation
• automated task generation
• trajectory augmentation
• vision-language-action policies