Despite recent progress in AI, deploying robots is still a time-consuming and cumbersome process. This is because current robots can only repeat pre-programmed instructions with minimum adaptability to changes. In this report, we introduce the concept of Teleported AI that draws from state-of-the-art deep neural networks, robot simulations and cloud-based software development to enable the massively parallel training of robots in virtual environments. This novel approach alleviates the need for expensive physical robot setups and makes progress in computing technology directly available to robotics with the goal of developing robots with adaptive autonomous behavior through simple direct instruction.
«
Despite recent progress in AI, deploying robots is still a time-consuming and cumbersome process. This is because current robots can only repeat pre-programmed instructions with minimum adaptability to changes. In this report, we introduce the concept of Teleported AI that draws from state-of-the-art deep neural networks, robot simulations and cloud-based software development to enable the massively parallel training of robots in virtual environments. This novel approach alleviates the need for...
»