Neural Network

What is a neural network?

• "A type of machine learning model"
• "A hardware component"
• "A type of data storage"
• "A programming language"

Answer explanation

"A neural network is a type of machine learning model designed to recognize patterns inspired by the structure of the human brain."

What is the primary function of the activation function in a neural network?

• "To initialize weights"
• "To transform input data"
• "To add non-linearity"
• "To determine the output layer"

Answer explanation

"The activation function introduces non-linearity into the model allowing it to learn complex patterns."

What does the term 'epoch' refer to in training a neural network?

• "A single pass through the data"
• "The final output of the model"
• "The architecture of the network"
• "The size of the dataset"

Answer explanation

"An epoch is one complete iteration through the entire training dataset."

What is backpropagation?

• "A method for weight initialization"
• "A technique for model evaluation"
• "An optimization algorithm"
• "A process for adjusting weights based on error"

Answer explanation

"Backpropagation is the algorithm used to minimize the error by adjusting the weights of the network based on the gradient of the loss function."

What is a convolutional neural network (CNN) primarily used for?

• "Text processing"
• "Time series prediction"
• "Image recognition"
• "Reinforcement learning"

Answer explanation

"CNNs are specifically designed for processing and analyzing visual data making them highly effective for image recognition tasks."

What role do neurons play in a neural network?

• "They store data"
• "They perform mathematical operations"
• "They manage the training process"
• "They control the hardware"

Answer explanation

"Neurons in a neural network compute outputs based on input data and weights applying an activation function."

What does the term 'overfitting' mean?

• "Model performs well on training data but poorly on test data"
• "Model is too simple"
• "Model has too many layers"
• "Model generalizes well"

Answer explanation

"Overfitting occurs when a model learns the noise in the training data leading to poor generalization to new data."

What is dropout in neural networks?

• "A type of optimization"
• "A regularization technique"
• "A layer type"
• "A loss function"

Answer explanation

"Dropout is a regularization technique used to prevent overfitting by randomly deactivating a subset of neurons during training."

What does a 'fully connected layer' mean in a neural network?

• "All neurons are connected to the next layer"
• "No neurons are connected"
• "Only some neurons are connected"
• "Neurons are connected only to the output"

Answer explanation

"In a fully connected layer every neuron from the previous layer is connected to every neuron in the next layer."

What is the purpose of the loss function in a neural network?

• "To adjust weights"
• "To evaluate model performance"
• "To initialize parameters"
• "To determine the model architecture"

Answer explanation

"The loss function measures how well the model's predictions match the actual outcomes guiding the training process."