8/27/2017 Coursera | Online Courses From Top Universities.

Join for Free | Coursera

Hyperparameter tuning, Batch Normalization, Programming

10/10 points
Frameworks (100%)
Quiz, 10 questions

 Congratulations! You passed! Next Item

1/1
 points

1. 
If searching among a large number of hyperparameters, you should try values in a grid
rather than random values, so that you can carry out the search more systematically and not
rely on chance. True or False?

True

False

Correct 

1/1
 points

2. 
Every hyperparameter, if set poorly, can have a huge negative impact on training, and so all
hyperparameters are about equally important to tune well. True or False?

True

False

Correct 
Yes. We've seen in lecture that some hyperparameters, such as the learning rate, are
more critical than others.

1/1
 points

3. 
https://www.coursera.org/learn/deep-neural-network/exam/CzYDo/hyperparameter-tuning-batch-normalization-programming-frame… 1/6
8/27/2017 Coursera | Online Courses From Top Universities. Join for Free | Coursera

3. 
Hyperparameter tuning, Batch
During hyperparameter Normalization,
search, whether you try to babysitProgramming
one model (“Panda” strategy) or
10/10 points
Frameworks
train a lot of models in parallel (“Caviar”) is largely determined by: (100%)
Quiz, 10 questions
Whether you use batch or mini-batch optimization

The presence of local minima (and saddle points) in your neural network

The amount of computational power you can access

Correct 

The number of hyperparameters you have to tune

1/1
 points

4. 
If you think β (hyperparameter for momentum) is between on 0.9 and 0.99, which of the
following is the recommended way to sample a value for beta?

1 r = np.random.rand()
2 beta = r*0.09 + 0.9

1 r = np.random.rand()
2 beta = 1-10**(- r - 1)

Correct 

1 r = np.random.rand()
2 beta = 1-10**(- r + 1)

1 r = np.random.rand()
2 beta = r*0.9 + 0.09

1/1
 points

5. 
https://www.coursera.org/learn/deep-neural-network/exam/CzYDo/hyperparameter-tuning-batch-normalization-programming-frame… 2/6
8/27/2017 Coursera | Online Courses From Top Universities. Join for Free | Coursera

5. 
Hyperparameter
Finding goodtuning, Batch
hyperparameter Normalization,
values Programming
is very time-consuming. So typically you should do it
10/10 points
Frameworks
once at the start of the project, and try to nd very good hyperparameters so that you don’t
(100%)
ever have to revisit tuning them again. True or false?
Quiz, 10 questions

True

False

Correct 

1/1
 points

6. 
In batch normalization as presented in the videos, if you apply it on the lth layer of your
neural network, what are you normalizing?

[
W l]

[
z l]

Correct 

[
a l]

[
b l]

1/1
 points

7. 
(i)
z −μ
In the normalization formula znorm , why do we use epsilon?
(i)
=
√σ 2 +ε

To avoid division by zero

Correct 

To speed up convergence

https://www.coursera.org/learn/deep-neural-network/exam/CzYDo/hyperparameter-tuning-batch-normalization-programming-frame… 3/6
8/27/2017 Coursera | Online Courses From Top Universities. Join for Free | Coursera

In case μ is too small

Hyperparameter tuning, Batch Normalization, Programming 10/10 points
Frameworks To have a more accurate normalization (100%)
Quiz, 10 questions

1/1
 points

8. 
Which of the following statements about γ and β in Batch Norm are true?

There is one global value of γ ∈ R and one global value of β ∈ R for each layer,
and applies to all the hidden units in that layer.

Un-selected is correct 

They can be learned using Adam, Gradient descent with momentum, or RMSprop,
not just with gradient descent.

Correct 

They set the mean and variance of the linear variable z [ l] of a given layer.

Correct 

β and γ are hyperparameters of the algorithm, which we tune via random

sampling.

Un-selected is correct 

−−−−−
The optimal values are γ 2
= √σ + ε , and β = μ.

Un-selected is correct 

1/1
 points

9. 

https://www.coursera.org/learn/deep-neural-network/exam/CzYDo/hyperparameter-tuning-batch-normalization-programming-frame… 4/6
8/27/2017 Coursera | Online Courses From Top Universities. Join for Free | Coursera

After training a neural network with Batch Norm, at test time, to evaluate the neural network
Hyperparameter tuning,
on a new example Batch Normalization, Programming
you should:
10/10 points
Frameworks (100%)
Quiz, 10 questions Perform the needed normalizations, use μ and σ 2 estimated using an
exponentially weighted average across mini-batches seen during training.

Correct 

If you implemented Batch Norm on mini-batches of (say) 256 examples, then to

evaluate on one test example, duplicate that example 256 times so that you’re
working with a mini-batch the same size as during training.

Skip the step where you normalize using μ and σ 2 since a single test example
cannot be normalized.

Use the most recent mini-batch’s value of μ and σ 2 to perform the needed
normalizations.

1/1
 points

10. 
Which of these statements about deep learning programming frameworks are true? (Check
all that apply)

Deep learning programming frameworks require cloud-based machines to run.

Un-selected is correct 

Even if a project is currently open source, good governance of the project helps
ensure that the it remains open even in the long term, rather than become closed
or modi ed to bene t only one company.

Correct 

A programming framework allows you to code up deep learning algorithms with

typically fewer lines of code than a lower-level language such as Python.

Correct 

https://www.coursera.org/learn/deep-neural-network/exam/CzYDo/hyperparameter-tuning-batch-normalization-programming-frame… 5/6
8/27/2017 Coursera | Online Courses From Top Universities. Join for Free | Coursera

Hyperparameter tuning, Batch Normalization, Programming 10/10 points

Frameworks (100%)
Quiz, 10 questions
  

https://www.coursera.org/learn/deep-neural-network/exam/CzYDo/hyperparameter-tuning-batch-normalization-programming-frame… 6/6