NVIDIA is now developing robots that are capable of learning actions from human beings. Although at initial stages, the development seems quite promising for Machine Learning as well as the implementation of bots in regular scenarios. The idea is to extend the development into a scenario where robots can be used to do certain tasks in industrial scenarios as well as home scenarios. This will put away the need of creating specific robots. Instead, standard robots can be made to learn the actions from human beings and imitate actions in a better manner. A lot of technologies are used to make this possible.
The latest reports are from the lab environment of NVIDIA, where the robots were capable of moving boxes or certain color as well as a toy car. In the course of time, the robots can be taught to do complex actions as well. NVIDIA wants to address one of the major problems of using robots in industrial scenarios — inability to adapt to the major changes in the environment. When developing a robot that has learning capability, it can be removed wisely. In most of the current cases, the core program of the robot will be changed, and it hasn’t been so affordable for the common people.
Using this technology, NVIDIA will make it possible for the common user to show a new task for the robot — following which the iteration would be done by the robot. From the very act of perception to the complex part of execution, these robots will be relying on neural networks, which are powered by AI and ML. This is also marking the changes in the field of robotics development, since experts are now trying to make robots more amiable to human interaction instead of following mundane commands. NVIDIA has not shown more details about its completion date or consumer-level implementation either.