Practice Activation Function (relu) (23.4.3) - Convolutional Neural Network (CNN)

Practice: Activation Function (ReLU)

Lesson 23.4.3Convolutional Neural Network (CNN)

Practise this section

9 questions

Flashcard drill
3 cards from this section. Quick, good the night before a test.
Exercise
6 questions · about 4 min · pass mark 70%. Hints available.
Sign up to start the exercise
Quiz
3 questions · about 2 min · pass mark 70%. Hints available.
Sign up to start the quiz

6 sample questions

Question 1.
What does the ReLU activation function do?
Show hintHide hint
Think about how it handles negative numbers.
Question 2.
Why is ReLU preferred over sigmoid in neural networks?
Show hintHide hint
Consider how gradients affect the learning process.
3Question 3.
What does the ReLU activation function do?
- A. Replaces all values with zero
- B. Replaces negative values with zero
- C. Keeps all values unchanged
Show hintHide hint
Remember the main function of ReLU.
4Question 4.
True or False: ReLU is only used in the output layer of a neural network.
- True
- False
Show hintHide hint
Think about where ReLU is applied in the architecture.
Question 5.
Imagine you are training a CNN and notice some neurons consistently output zero. What measures can you implement to improve this situation?
Show hintHide hint
What alternative activation functions could you utilize?
Question 6.
Discuss the impact of ReLU activation on a neural network's performance when dealing with multi-class classification tasks involving sparse data.
Show hintHide hint
Think about how multi-class tasks can involve varied feature ranges.

5 more questions in this section. Enroll free to practise them all.