The Open X-Embodiment database features more than 500 skills and 150,000 tasks from 22 different robot types. The creators are making the data available to the research community, hoping it will reduce barriers and accelerate research. The future of robotics, according to the team, relies on enabling robots and researchers to learn from each other.
Key takeaways:
- The transition from single to general purpose robotics is a major challenge in the industry, with robot learning being a key solution.
- Many robotics research labs, startups, and corporations are working on this issue, with a need for a large, shared dataset being central to most solutions.
- Google’s DeepMind robotics team has collaborated with 33 research institutes to create a massive, shared database called Open X-Embodiment, which is likened to ImageNet in the field of computer vision.
- The Open X-Embodiment database, which features more than 500 skills and 150,000 tasks from 22 different robot types, is being made available to the research community to accelerate research and enable robots and researchers to learn from each other.