Module 1 - Introduction to Deep Learning

1. How are Traditional Machine Learning and Deep Learning different when it comes to data needs?

Deep Learning needs a lot of data yet Traditional Machine Learning needs much fewer data points to train.

1. How many types of layers do neural networks have? And what are they called?

Neural networks have 3 types of layers: input, output and hidden layer.

1. What Deep Learning technique is commonly used for computer vision?

CNNs are good at computer vision tasks.

Module 2 - Building Blocks of Deep Learning

1. What are the two main parameters that are learned?

Weight and bias are the two main parameters that are learned during training.

1. Why do we calculate things in a vectorized way?

To be able to calculate the output of the network simultaneously for all the data points. This makes the network faster.

1. Why do we need a non-linear activation function?

Because otherwise, the network is just a linear transformation of the inputs. And no matter how big the network gets it will be as good as a single neuron because of this linearity. This will cause the network to not be able to fit complex data.

1. What is the goal of backpropagation?