Skip to main content

New DeepMind AI model can control robotic actions it was never trained to do


Google DeepMind has a new AI model that can direct robotic tasks it was never trained to perform. Named RT-2, the model learns from web and robotics data. It then turns this information into simple instructions for machines. In tests, the model was asked to take actions never seen in the robotic data, such as placing oranges in a matching bowl. To follow these commands, the system had to translate knowledge from web-based data. According to DeepMind, the model had a 62% success for these operations — double that of its predecessor, RT-1. “Just like language models are trained on…

This story continues at The Next Web https://ift.tt/ZzxaOWR
Read full article: The Next Web
Labels:

Comments

Post a Comment

What's new?

Popular posts from this blog

My credit

February 05, 2014 at 10:23AM: This app offers detailed information on how to repair your own credit, it\'s function is only to provide information on how to do it by sending letters to the three credit reporting agencies, this app is to help those that want to correct negative information from their credit reports, without having to use the services of a professional credit repair agency, though the option of doing so is also available with the owner of this app. Read full article: FeedMyApp
Post a Comment
Read more