Cornell researchers built an artificial neural network for self-driving cars to make the cars remember routes and signs; each vehicle that uses this network can identify what constitutes traffic participants and what is safe to ignore. After several tests runs on the road, these vehicles can reliably detect and identify objects, even if they are traveling on a new route.
Cornell researchers presented this research at the Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR 2022), last June, in New Orleans.