5 Neural Networks Free Online FlashCards

What is a neural network?

A machine-learning model made of connected computational units arranged in layers that learns patterns by adjusting numerical parameters.

What does the input layer do?

It receives the example’s features, such as image pixel values or measurements in a table.

What happens during a forward pass?

A forward pass sends an input through the network layer by layer to make a prediction.

How can a validation set help assess a network?

A separate validation set helps check whether learned patterns generalize to new examples and can reveal overfitting.

What kinds of tasks use neural networks?

They are used for image and video analysis, language tasks, speech recognition, prediction, and pattern detection.

When are neural networks especially useful?

They can learn complex relationships in data that are difficult to describe with a simple formula or hand-written rules.

How do hidden-layer units transform inputs?

They transform inputs by combining values from the preceding layer using weights and a bias.

What does the output layer produce?

It produces the model’s prediction, such as a category or a numerical estimate.

Why do neural networks use activation functions?

An activation function applies a nonlinear transformation to a unit’s result. Without nonlinear activations, stacking layers would still produce only a linear relationship.

Which values are a neural network’s parameters?

The learned weights and biases are the network’s parameters.

What does a loss function measure?

A loss function compares a prediction with the correct answer and measures the prediction error.

What does backpropagation calculate?

Backpropagation calculates how changing each parameter would affect the loss.