RNN, LSTM, and CNN

What does CNN stand for in Deep Learning?

• Central Neural Network

• Convolutional Neural Network

• Complex Neural Network

• Combined Neural Network

What is the main purpose of convolution layers in CNN?

• To reduce overfitting

• To extract features from input

• To store memory

• To normalize data

Which type of data is CNN mainly used for?

• Tabular data

• Time-series data

• Image data

• Text-only data

What does a kernel/filter do in CNN?

• Adds noise to the image

• Extracts local features

• Stores weights of the model

• Increases dataset size

Pooling layers in CNN are mainly used for:

• Increasing resolution

• Feature extraction

• Reducing spatial dimensions

• Converting image to grayscale

Which pooling method is most commonly used in CNNs?

• Average pooling

• Max pooling

• Min pooling

• Median pooling

The output of convolution operation is called:

• Feature map

• Hidden state

• Gradient map

• Vector space

Why are CNNs better than fully connected networks for images?

• They use fewer parameters

• They train faster always

• They don’t need activation functions

• They never overfit

What is stride in convolution operation?

• Size of filter

• Number of filters used

• Step size of filter movement

• Depth of convolution

Which activation function is most commonly used in CNN hidden layers?

• Sigmoid

• Tanh

• ReLU

• Softmax