Motional and MIT develop explainable system for autonomous driving
Motional and researchers from MIT’s Computer Science and Artificial Intelligence Laboratory have developed a system intended to show why a self-driving vehicle makes particular decisions. The work, published in Nature, is led by a Motional team that includes CEO Laura Major. Their method, called the Concept-Wrapper Network, or CW-Net, converts neural-network activity into readable concepts such as “Approaching Stopped Vehicle” and “Close to Cyclist.”