DL-ANN

The basic unit of an Artificial Neural Network is called:

• Kernel

• Perceptron

• Decision Tree

• Support Vector

The activation function in ANN is used to:

• Scale the input

• Normalize data

• Introduce non-linearity

• Measure accuracy

The purpose of using multiple hidden layers in deep learning is to:

• Increase model simplicity

• Reduce computation

• Eliminate overfitting

• Learn complex hierarchical patterns

The process of adjusting weights in ANN during training is called (a)  
(answer in small case letters)

Which loss function is commonly used in binary classification with ANN?

• Cross-Entropy Loss

• Hinge Loss

• Gini Index

• Mean Squared Error

The number of neurons in the input layer of an ANN corresponds to:

• Number of classes

• Number of features

• Number of hidden layers

• Number of epochs

The main advantage of deep learning over traditional machine learning is:

• Does not require GPUs

• Always faster to train

• Automatically learns features

• Requires less data

ReLU activation function outputs:

• between 0 and 1

• Negative values only

• Max(0, x)

• Always binary outputs

  In deep learning, overfitting can be reduced by:

• Data Augmentation

• Adding more layers

• Increasing the learning rate

• Removing dropout layers

The learning rate in ANN controls:

• Step size during weight update

• Number of hidden layers

• Size of the dataset

• Number of neurons