State-of-the-Art Technology 2018-10-25T21:22:39+00:00

State-of-the-Art Technology

Our new approach for training large, unlabeled data-sets with minimal reward engineering allows us to collect data from both real-world and simulation examples. Manual labeling imagery algorithms train the driving model to control the car by extracting the most useful driving-parameters and abstract models of the road, allowing us to leverage data collected by our simulators.