Robot learning has advanced tremendously in the last decade. From learning low-level manipulation skills to long-horizon mobile manipulation tasks and autonomous driving, machine learning has accelerated the advancement in the entire spectrum of robotic domains. Much of this success has been fueled by data-driven learning algorithms, massive, curated datasets, and the doubling of computational capacity each year. We also witness more and more learned robotic systems performing tasks in human- centered environments alongside humans. Notable areas include robots in collaborative manufacturing, agriculture, logistics, and search and rescue operations.

中文翻译：

---

 扫描问题


机器人学习在过去十年中取得了巨大进步。从学习低级操作技能到长期移动操作任务和自动驾驶，机器学习加速了整个机器人领域的进步。这种成功很大程度上是由数据驱动的学习算法、海量、精心策划的数据集以及每年翻倍的计算能力推动的。我们还见证了越来越多的学习机器人系统在以人为中心的环境中与人类一起执行任务。值得注意的领域包括协作制造、农业、物流以及搜救行动中的机器人。