ARIO (All Robots In One)
收藏资源简介:
ARIO数据集由南方科技大学、中山大学和鹏城实验室联合创建,旨在为多用途、通用型具身智能代理提供标准化的数据格式。该数据集包含约300万条记录,涵盖258个系列和321,064个任务,结合了真实世界和模拟数据。创建过程中,数据集通过多平台收集、模拟生成和开源数据转换等方式构建。ARIO数据集的应用领域广泛,主要用于提高具身智能代理的鲁棒性和适应性,解决数据格式不统一、多样性不足和数据量不足等问题。
The ARIO dataset was jointly developed by Southern University of Science and Technology, Sun Yat-sen University, and Peng Cheng Laboratory, with the aim of providing standardized data formats for versatile, general-purpose embodied AI agents. This dataset contains approximately 3 million records, covering 258 series and 321,064 tasks, and combines real-world and simulated data. During its development, the dataset was constructed through multiple approaches including multi-platform data collection, simulated data generation, and open-source data transformation. The ARIO dataset has a wide range of application scenarios, and is mainly used to improve the robustness and adaptability of embodied AI agents, addressing issues such as inconsistent data formats, insufficient data diversity, and inadequate data volume.




