Optimizing Embedded Vision Efficiency across the Hardware-dataflow-algorithm Stack KU Leuven
Artificial Intelligence (AI) systems have been shown to be capable of performing complex tasks like (semi-)autonomous driving, image enhancement or creation, and winning strategy games such as Chess and even Go. All these example of AI are powered by a particularly powerful technique: neural networks (NN). However, processing neural networks is very demanding of the hardware executing them. Yet, at the same time, they should run on ...