Google DeepMind Revolutionizes Robotics with Gemini Robotics-ER 1.6
Le brief IA que les pros lisent chaque soir
Les 7 actus IA du jour, décryptées en 5 min. Gratuit.
Inclus dès l'inscription : notre sélection des meilleurs guides & comparatifs IA.
Choisis ton rythme
Gratuit · Pas de spam · Désabonnement en 1 clic
Google Deepmind recently introduced Gemini Robotics-ER 1.6, a model for embodied reasoning in robots. This model serves as a sophisticated reasoning layer, enabling robots to better understand their environment and autonomously plan tasks. To achieve this, it utilizes tools such as Google Search and models that integrate vision, language, and action.
According to Deepmind, Gemini Robotics-ER 1.6 surpasses previous versions, including Gemini Robotics-ER 1.5 and Gemini 3.0 Flash, in object designation, counting, and recognizing successful task execution. This improvement is crucial for the development of smarter and more autonomous robots.
In collaboration with Boston Dynamics, the model's ability to read instruments such as gauges and level glasses has been significantly enhanced. The model combines agentic image processing with code execution, allowing for a closer look at small display details. It then uses designation functions and code to calculate proportions and assess distances before applying world knowledge to interpret the readings. Boston Dynamics' Spot robot could utilize this feature to perform system inspections with increased accuracy.
The Gemini Robotics-ER 1.6 model is available via the Gemini API and Google AI Studio, along with a Colab example for developers.
Brief IA — L'actualité IA en français
L'essentiel de l'actualité de l'intelligence artificielle, décrypté et expliqué chaque jour.