Weight initialization strategies in TensorFlow - Cheat Sheet & Quick Revision

Recall & Review

beginner

What is the purpose of weight initialization in neural networks?

Weight initialization sets the starting values of the model's weights before training. Good initialization helps the model learn faster and avoid problems like vanishing or exploding gradients.

intermediate

Explain the difference between Xavier (Glorot) and He initialization.

Xavier initialization sets weights to keep variance of activations similar across layers, good for sigmoid/tanh activations. He initialization adjusts weights for ReLU activations to better handle their behavior and avoid dying neurons.

beginner

What problem does He initialization help to solve?

He initialization helps prevent the problem of dying ReLU neurons by scaling weights properly so that the variance of outputs stays stable through layers.

beginner

How does random initialization differ from zero initialization, and why is zero initialization usually bad?

Random initialization assigns small random values to weights, breaking symmetry so neurons learn different features. Zero initialization sets all weights to zero, causing all neurons to learn the same thing and preventing effective training.

beginner

Name two common TensorFlow functions used for weight initialization.

Two common TensorFlow initializers are tf.keras.initializers.GlorotUniform() for Xavier initialization and tf.keras.initializers.HeNormal() for He initialization.

Which weight initialization method is best suited for ReLU activation functions?

A He initialization

B Xavier initialization

C Zero initialization

D Random initialization with zeros

He initialization is designed to work well with ReLU activations by scaling weights to keep variance stable.

Why is zero initialization of weights usually a bad idea?

A It makes the model too complex

B It causes all neurons to learn the same features

C It uses too much memory

D It speeds up training too much

Zero initialization causes symmetry where all neurons update identically, preventing the network from learning diverse features.

What does Xavier initialization aim to keep stable across layers?

A Number of neurons

B Learning rate

C Variance of activations

D Bias values

Xavier initialization keeps the variance of activations stable to help gradients flow well during training.

Which TensorFlow initializer corresponds to Xavier initialization?

A tf.keras.initializers.Zeros()

B tf.keras.initializers.HeNormal()

C tf.keras.initializers.RandomNormal()

D tf.keras.initializers.GlorotUniform()

GlorotUniform is TensorFlow's implementation of Xavier initialization.

What problem can occur if weights are initialized with very large values?

A Exploding gradients

B Vanishing gradients

C Faster convergence

D Better accuracy

Large initial weights can cause exploding gradients, making training unstable.

Describe why weight initialization matters and name two common strategies.

Think about how starting weights affect learning speed and stability.

Explain how He initialization helps with ReLU activations and what problem it prevents.

Consider the behavior of ReLU and how weights influence neuron output.

Practice

1. Which weight initialization strategy is best suited for layers using ReLU activation in TensorFlow?

easy

A. Zeros initializer

B. GlorotUniform initializer

C. RandomNormal initializer with mean 0

D. HeNormal initializer

Solution

1. Step 1: Understand activation and initialization relation

   ReLU activation benefits from initial weights that maintain variance through layers to avoid dying neurons.

2. Step 2: Match initializer to activation

   HeNormal initializer is designed for ReLU, scaling weights to keep variance stable.

3. Final Answer:

   HeNormal initializer -> Option D

4. Quick Check:

   ReLU uses HeNormal [OK]

Hint: Use HeNormal for ReLU activations to keep variance stable [OK]

Common Mistakes:

• Choosing GlorotUniform for ReLU layers
• Using zeros initializer causing no learning
• RandomNormal without scaling causes slow training

2. Which of the following is the correct syntax to initialize weights with GlorotUniform in TensorFlow?

easy

A. tf.keras.initializers.glorot_uniform()

B. tf.initializers.GlorotUniform()

C. tf.keras.initializers.GlorotUniform()

D. tf.keras.initializers.Glorot_uniform()

Solution

1. Step 1: Recall TensorFlow initializer module

   Initializers are accessed via tf.keras.initializers with exact class names.

2. Step 2: Check correct class name and casing

   The correct class name is GlorotUniform with capital G and U, called as a class with parentheses.

3. Final Answer:

   tf.keras.initializers.GlorotUniform() -> Option C

4. Quick Check:

   Correct class name and module path [OK]

Hint: Use exact class name with correct casing and parentheses [OK]

Common Mistakes:

• Using lowercase function names instead of class
• Missing parentheses after class name
• Using wrong module path tf.initializers

3. What will be the output shape and initializer type of the following TensorFlow layer?

layer = tf.keras.layers.Dense(10, kernel_initializer=tf.keras.initializers.HeNormal())
print(layer.kernel_initializer)

medium

A. GlorotUniform initializer object, output shape (None, 10)

B. HeNormal initializer object, output shape (None, 10)

C. HeNormal initializer object, output shape (10, 10)

D. RandomNormal initializer object, output shape (None, 10)

Solution

1. Step 1: Identify initializer and layer output

   The layer uses HeNormal initializer explicitly for kernel weights and has 10 output units.

2. Step 2: Understand output shape of Dense layer

   Dense layers output shape is (batch_size, units), so shape is (None, 10) before input is given.

3. Final Answer:

   HeNormal initializer object, output shape (None, 10) -> Option B

4. Quick Check:

   HeNormal used and output units = 10 [OK]

Hint: Dense output shape is (None, units); initializer matches given class [OK]

Common Mistakes:

• Confusing output shape with kernel shape
• Assuming default initializer instead of HeNormal
• Mixing initializer types

4. You wrote this code to initialize weights but get an error:

layer = tf.keras.layers.Dense(5, kernel_initializer='he_normal')
model = tf.keras.Sequential([layer])
model.build(input_shape=(None, 3))
print(layer.kernel_initializer())

What is the cause of the error?

medium

A. Calling kernel_initializer as a function without arguments

B. Using string 'he_normal' instead of initializer object

C. Not specifying bias_initializer

D. Missing input_shape in layer constructor

Solution

1. Step 1: Understand kernel_initializer attribute type

   kernel_initializer is an initializer object, not a callable function without arguments.

2. Step 2: Identify error cause

   Calling kernel_initializer() without arguments causes error because it expects shape and dtype parameters.

3. Final Answer:

   Calling kernel_initializer as a function without arguments -> Option A

4. Quick Check:

   Initializer needs shape args, can't call empty () [OK]

Hint: Don't call kernel_initializer() without shape and dtype [OK]

Common Mistakes:

• Confusing string initializer with object
• Ignoring required arguments for initializer call
• Assuming bias_initializer causes this error

5. You want to initialize weights for a model with mixed activations: ReLU in some layers and sigmoid in others. Which strategy best improves training speed and accuracy?

hard

A. Use HeNormal for ReLU layers and GlorotUniform for sigmoid layers

B. Use GlorotUniform for all layers regardless of activation

C. Use RandomNormal with mean 0 and stddev 1 for all layers

D. Use Zeros initializer for all layers to start from zero

Solution

1. Step 1: Match initializers to activation functions

   HeNormal suits ReLU to maintain variance; GlorotUniform suits sigmoid/tanh to avoid saturation.

2. Step 2: Understand impact on training

   Proper matching prevents slow or stuck training and improves accuracy by keeping gradients healthy.

3. Final Answer:

   Use HeNormal for ReLU layers and GlorotUniform for sigmoid layers -> Option A

4. Quick Check:

   Match initializer to activation for best training [OK]

Hint: Match HeNormal to ReLU, GlorotUniform to sigmoid/tanh [OK]

Common Mistakes:

• Using same initializer for all activations
• Using zeros initializer causing no learning
• Using RandomNormal without scaling