Researchers have developed AI models, called robot utility models (RUMs), that enable robots to perform basic tasks in new environments without needing further training. The RUMs allow machines to complete five separate tasks in unfamiliar environments with a 90% success rate, potentially making it quicker and easier to teach robots new skills and deploy them in homes. Teaching robots new skills typically requires a lot of data, which can be time-consuming and expensive to collect physically. To speed up the process, researchers used a setup to record demonstrations in 40 different home environments and trained learning algorithms on the data sets to create the RUM models. The researchers then deployed the models on a robot to test their success rate in executing tasks in new environments. While the completion rate initially stood at 74.4%, it increased to 90% when the researchers used images from the robot’s camera and an AI model to evaluate task completion. This research is significant because it helps robots behave more reliably in new settings, and it could serve as a recipe to build other utility robotics models for different tasks. The ultimate goal is to make it easier for untrained individuals to deploy future robots in their homes.
The ainewsarticles.com article you just read is a brief synopsis; the original article can be found here: Read the Full Article…
