Helping Robots Understand What Things Are
Humans are very adaptable.
Humans have the ability to understand the various categories of things in the world. That understanding is then used to deal with things that a person has never encountered before. That’s just one of the things which makes humans different from robots.
For example, no one has seen all the different types of doors that are in use all over the world, but when a person sees a new door for the first time, he can tell what it is, based on what he knows about doors in general.
How Robots Understand the World
The level of understanding that humans have of the world is lacking in most robots.
Some experts argue that robots will have to come to an understanding of the world that is close to the human level if they are going to be assigned to completely autonomous work or work that is independent from human intervention.
Making Robots Understand Things
During the 2019 Conference on Computer Vision and Pattern Recognition, a group of researchers announced the PartNet.
So, what is PartNet and how can it help in developing robotic technology? PartNet is a database of 3D objects that have been annotated down to the level which can be used by robots to understand what those objects are in the same level that humans do.
PartNet is actually just a subset of an even bigger database known as ShapeNet. What makes PartNet different is the level of annotation. It’s not just a database of images but also information about each object, making it easy for AIs or robots to develop a more complete understanding of that object.
It took a tremendous amount of work to annotate the over 26,000 3D shapes in the database and it’s just the start. Researchers hope PartNet will be the beginning of a system that can annotate 3D models on its own, making it an autonomous AI itself.
Just imagine how awesome robots will be once they understand the objects around them on their own. The robot vacuums of iClebo can already do so much when it comes to navigation and figuring out how to move around a room, imagine what they can do once they can understand their surroundings better.