Neural Networks and SVM Quiz

What is the primary function of a multi-layer perceptron (MLP)?

• To classify data into predefined categories

• To reduce the dimensionality of data

• To solve clustering problems

• To find anomalies in data

In a multi-layer perceptron, what is the purpose of the hidden layer?

• To output the final prediction of the model

• To transform inputs into a format that can reveal complex patterns

• To directly connect input and output layers

• To prevent overfitting

The process of updating weights in a neural network by minimizing error through backward propagation is known as:

• Forward propagation

• Weight initialization

• Backpropagation

• Regularization

Which of the following best describes backpropagation in neural networks?

• A method to add new layers to the neural network

• An algorithm for calculating the gradient of the loss function to update weights

• A regularization technique to prevent overfitting

• A method for encoding inputs in high-dimensional space

In a multi-layer perceptron, how is the error propagated back through the layers?

• By calculating the partial derivatives of the error with respect to each weight

• By initializing weights to random values

• By using a clustering algorithm

• By applying the linear regression formula

Which activation function is commonly used in hidden layers of MLPs to introduce non-linearity?

• Linear

• Step function

• Sigmoid or ReLU (Rectified Linear Unit)

• Threshold function

What is the purpose of forward propagation in an MLP?

• To propagate error back through the network

• To initialize weights

• To calculate the output based on input data and current weights

• To normalize the input data

In the context of MLPs, what is a 'loss function'?

• A measure of the model's accuracy

• A function that evaluates the error of the network's predictions

• A method for regularizing the model

• A type of activation function

Which of the following is an example of a commonly used loss function for classification tasks in MLPs?

• Mean Squared Error

• Binary Cross-Entropy

• Root Mean Squared Error

• Mean Absolute Error

When using backpropagation, which method is commonly used to minimize the loss function?

• K-means clustering

• Gradient descent

• Random initialization

• Genetic algorithms