August 2019
Intermediate to advanced
202 pages
5h 9m
English
Gradient descent combined with backpropagation is the most popular learning algorithm in recent machine learning or deep neural network systems. There are three kinds of gradient descent:
Here, batch gradient descent is not practical in the majority of cases due to hardware memory limitations for large datasets. Also, stochastic gradient descent could be slow since a model learns from one piece of data at a time. Due to these reasons, mini-batch gradient descent is the most widely ...
Read now
Unlock full access