Google DeepMind hat zwei innovative KI-Modelle vorgestellt, die Robotern helfen, komplexe Aufgaben zu meistern.
In Kürze
- Gemini Robotics und Gemini Robotics-ER erleichtern Robotern das Leben.
- Vielseitigkeit: Anpassbar an verschiedene Roboterplattformen.
- Neues Sicherheitsframework für sichere Roboterhandlungen.
Google DeepMind’s New AI Models: Gemini Robotics and Gemini Robotics-ER
Google DeepMind has recently introduced two exciting AI models: Gemini Robotics and Gemini Robotics-ER. These new technologies are designed to make life easier for robots in the real world. They combine language processing, visual perception, and physical actions to independently master complex tasks such as folding origami or organizing a desk. With their enhanced spatial intelligence, these robots can accurately recognize and safely handle objects in space – an example of this is the safe grasping of a coffee cup.
Versatility and Adaptability
A particularly noteworthy aspect of these models is their versatility. They are not restricted to specific robot platforms but can adapt to various types – from industrial arms to humanoid robots. Google has also placed great emphasis on safety. A newly developed framework allows robot actions to be controlled in natural language, while the ASIMOV dataset helps increase the safety of the robots.
Collaborations and Future Prospects
The development of these models is in collaboration with renowned companies like Boston Dynamics. Initial tests show promising results, suggesting a wide range of applications in the future. The question remains how quickly and safely these robots can be integrated into our daily lives and industry. It will be exciting to observe what role they will play in the near future.
Quellen
- Quelle: Google DeepMind
- Der ursprüngliche Artikel wurde hier veröffentlicht
- Dieser Artikel wurde im Podcast KI-Briefing-Daily behandelt. Die Folge kannst du hier anhören.
