Google's DeepMind AI enables robots to perform novel tasks

Source: IANS

Google has demonstrated its first vision-language-action (VLA) model for robot control that showed improved generalization capabilities and semantic (the understanding of words and sentences), and visual understanding beyond the robotic data it was exposed to.

This includes interpreting new commands and responding to user commands by performing beginner-level reasoning, such as reasoning about object categories or high-level descriptions.

The Robotic Transformer 2 (RT-2) is a novel vision-language-action (VLA) model that learns from both web and robotics data, and translates this knowledge into generalized instructions for robotic control, according to Google DeepMind, the tech giant's artificial intelligence subsidiary.

A robot that can perform multiple tasks

A traditional robot can pick up a ball and stumble when picking up a cube. RT-2's flexible approach enables a robot to train on picking up a ball and can figure out how to adjust its extremities to pick up a cube or another toy it's never seen before.

"We also show that incorporating chain-of-thought reasoning allows RT-2 to perform multi-stage semantic reasoning, like deciding which object could be used as an improvised hammer (a rock), or which type of drink is best for a tired person (an energy drink),” said the team behind it.

The latest model builds upon Robotic Transformer 1 (RT-1) that was trained on multi-task demonstrations.

The team performed a series of qualitative and quantitative experiments on RT-2 models, on over 6,000 robotic trials.

The potential of vision-language models

The RT-2 model shows that vision-language models (VLMs) can be transformed into powerful vision-language-action (VLA) models, which can directly control a robot by combining VLM pre-training with robotic data.

"RT-2 is not only a simple and effective modification over existing VLM models, but also shows the promise of building a general-purpose physical robot that can reason, problem solve, and interpret information for performing a diverse range of tasks in the real -world,” said Google DeepMind.

(Source: IANS)

-Edited for style

For more such updates and news on the go, follow us on

Instagram | YouTube | Facebook

Comments

Popular posts from this blog

Nokia wins case against Oppo and OnePlus, can't sell phones in this country

For the first time, NASA is sending a drone helicopter to Mars

Five decades after man's arrival, 'moon trip' difficult, 60 out of 146 missions fail

If you are interested in astronomy due to solar eclipse, let us know about an app to install in your phone.

Voice typing will become easier and faster

Is a solar storm moving towards the earth at a speed of 12 lakh km per hour?

Solid evidence of flowing water on Mars, the red planet of the Solar System: Images of pebbles and rocks found

Impact of global warming: The country's last 4,000-year-old iceberg also collapsed

Successful landing of private company Intuitive's Odysseus spacecraft on the South Pole of the Moon

The government tightened the grip on social media even more