Google’s artificial intelligence division DeepMind has launched a new model that allows powerful AI robots to operate without an internet connection.
Typically AI-powered robots rely on far-away servers to carry out the necessary computing, with DeepMind’s new Gemini Robotics On-Device model marking a major shift to this approach.
The on-device system allows humanoid robots to operate in virtually any location on Earth – or even in space – without the internet.
“Since the model operates independent of a data network, it’s helpful for latency sensitive applications, and ensures robustness in environments with intermittent or zero connectivity,” Carolina Parada, DeepMind’s head of robotics, wrote in a blog post.
“Gemini Robotics On-Device marks a step forward in making powerful robotics models more accessible and adaptable... We’re excited to see what the robotics community will build with these new tools as we continue to explore the future of bringing AI into the physical world.”
The next-generation artificial intelligence allows robots to independently carry out everyday tasks, like folding clothes or unzipping bags.
Dr Parada also noted that the on-device model is also able to quickly learn new tasks when shown between 50 to 100 demonstrations of how to perform it.
The model can also be adapted to work on different types of robots in order to carry out different types of tasks.
Other leading AI firms have also been working on this type of model for robotics, with Nvidia recently unveiling its Isaac Groot N1 model in an effort to supercharge humanoid robot development.
The “fast-thinking action model” is designed to mirror human reflexes and intuition in order to fill roles typically carried out by humans.
“The age of generalist robotics is here,” Jensen Huang, founder and CEO of Nvidia, said at the unveiling of the model in March.
“With Nvidia Isaac GR00T N1 and new data-generation and robot-learning frameworks, robotics developers everywhere will open the next frontier in the age of AI.”
