Algorithm and Method to Create a Functional Machine Learning Model to Be Used In An Automotive ECU

摘要

This invention describes an algorithm and method to create a functional machine learning model to be used in an Automotive ECU. The proposed algorithm for Machine Learning is Backpropagation. It is proposed to combine Backpropagation with a learning algorithm such as Gradient descent, in which we compute the "error" or "cost" function obtained by Backpropagation and then it is minimized until get the lowest failure rate possible as for many training examples as needed.