Yet, it is still common to solve these problems by writing complex logical algorithms. This presentation will showcase how Neonode applies the unique possibilities of machine learning in its driver and in-cabin monitoring system MultiSensing®. Utilizing proprietary synthetic training data sets, we have trained networks which provide fast, accurate and robust answers to common questions in driver monitoring scenarios. In this session, you will learn more about – Which annotations to choose for the training data of neural networks – Creating neural networks which are independent of camera type and position – How neural networks can solve complex problems in driver monitoring applications