Dense (fully connected) layers in TensorFlow - Cheat Sheet & Quick Revision

Recall & Review

beginner

What is a Dense (fully connected) layer in a neural network?

A Dense layer connects every input neuron to every output neuron with its own weight. It helps the model learn complex patterns by combining all inputs.

beginner

How does a Dense layer transform its input?

It multiplies the input by a weight matrix, adds a bias vector, then applies an activation function to produce the output.

beginner

Why do we use activation functions after Dense layers?

Activation functions add non-linearity, allowing the network to learn complex patterns beyond simple straight lines.

beginner

In TensorFlow, how do you add a Dense layer with 10 neurons and ReLU activation?

Use: tf.keras.layers.Dense(10, activation='relu') inside your model.

intermediate

What happens if you omit the activation function in a Dense layer?

The layer will perform only a linear transformation, limiting the model's ability to learn complex patterns.

What does a Dense layer do in a neural network?

A Connects every input to every output neuron

B Only connects inputs to outputs with zero weights

C Removes connections between neurons

D Only connects neurons in the same layer

A Dense layer fully connects every input neuron to every output neuron with weights.

Which operation is NOT part of a Dense layer's computation?

A Matrix multiplication of inputs and weights

B Adding bias terms

C Applying an activation function

D Pooling inputs

Pooling is a separate operation, not part of Dense layer computation.

Why do we add activation functions after Dense layers?

A To remove bias

B To reduce the number of neurons

C To add non-linearity so the model can learn complex patterns

D To make the layer linear

Activation functions add non-linearity, enabling complex pattern learning.

In TensorFlow, how do you create a Dense layer with 5 neurons and no activation?

A tf.keras.layers.Dense(activation='none')

B tf.keras.layers.Dense(5)

C tf.keras.layers.Dense(5, activation='none')

D tf.keras.layers.Dense(activation=None)

By default, Dense layers have no activation if not specified, so tf.keras.layers.Dense(5) creates 5 neurons with linear activation.

What is the role of the bias term in a Dense layer?

A To shift the output and help the model fit data better

B To multiply inputs

C To remove noise from data

D To connect neurons

Bias shifts the output, allowing the model to fit data more flexibly.

Explain in your own words what a Dense (fully connected) layer does in a neural network and why it is important.

Think about how every input influences every output and how the layer helps the model learn.

Describe how you would add a Dense layer in TensorFlow and what parameters you might specify.

Consider the basic syntax and common options like activation.