Real Time in Machine Learning

Real time can be very misleading when someone says real time in machine learning, because indeed it can be very, very different. For a bank, for example, a that is detecting fraudulent transactions, real time means probably a few minutes. But if you think about real time for a letsan autonymous vehicle that has to respond or take a decision upon the collection of data coming from censors or meras, or lidar cameras,. well, real time there is as a much stricter constraint in terms of time.