I just completed my Coursera course in Deep Learning Specialization!!

Courses offered:

1. Neural Networks and Deep Learning
2. Improving Deep Neural Networks: Hyperparameter tuning, Regulization, and Optimization
3. Structuring Machine Learning Projects
4. Convolutional Neural Networks
5. Sequence Models

These courses were very informative. One fundamental idea that one needs to remember is that neural networks are made up of a units of linear regression. And it was natural to start the course with the evaluation and examination of a linear regression. And from there scale it up to a single layer standard neural network, and scale it up more to invoke hidden layers. Deep neural network is just basically a neural network with more than two hidden layers.

Each course offers an effective message: “Machine learning is a very empirical process!” This implies the importance of experimentation on different parameters and hyperparameters. Ng stressed the importance of regularization and use of a different optimization algorithm both in weight updates and weight initialization.