exam questions

Exam DP-100 All Questions

View all questions & answers for the DP-100 exam

Exam DP-100 topic 6 question 1 discussion

Actual exam question from Microsoft's DP-100
Question #: 1
Topic #: 7
[All DP-100 Questions]

You need to implement a scaling strategy for the local penalty detection data.
Which normalization type should you use?

  • A. Streaming
  • B. Weight
  • C. Batch
  • D. Cosine
Show Suggested Answer Hide Answer
Suggested Answer: C 🗳️
Post batch normalization statistics (PBN) is the Microsoft Cognitive Toolkit (CNTK) version of how to evaluate the population mean and variance of Batch
Normalization which could be used in inference Original Paper.
In CNTK, custom networks are defined using the BrainScriptNetworkBuilder and described in the CNTK network description language "BrainScript."
Scenario:
Local penalty detection models must be written by using BrainScript.
Reference:
https://docs.microsoft.com/en-us/cognitive-toolkit/post-batch-normalization-statistics

Comments

Chosen Answer:
This is a voting comment (?). It is better to Upvote an existing comment if you don't have anything to add.
Switch to a voting comment New
huyennguyen
Highly Voted 4 years, 9 months ago
Both the question and answer are difficult to follow.
upvoted 78 times
...
satishgunjal
Highly Voted 3 years, 10 months ago
In case study they have also mentioned - All penalty detection models show inference phases using a Stochastic Gradient Descent (SGD) are running too slow - The images and videos will have varying sizes and formats So Batch normalization is usefull to speedup the process where as Cosine normalization is usefull to handle varying sizes and formats of input data.
upvoted 9 times
prashantjoge
3 years, 5 months ago
any reference for this?
upvoted 2 times
...
...
GHill1982
Most Recent 9 months, 3 weeks ago
Selected Answer: C
The best normalization type to use in this case is batch normalization. Batch normalization is a technique that reduces the internal covariate shift of the inputs to each layer of a neural network, making the training faster and more stable. Batch normalization also has the benefit of regularizing the model and reducing the need for dropout.
upvoted 1 times
...
snegnik
1 year, 5 months ago
Terminology from Cognitive Toolkit and Synapse Analytics. It seems doesn't relevant for DP-100 test
upvoted 1 times
...
ning
2 years, 4 months ago
DNN normalization?? I really do not expect this kind of questions ... The most common one is batch, and weight is kind of a batch with some improvements ... For other two, I do not know ...
upvoted 2 times
...
[Removed]
2 years, 6 months ago
is this question really for DP-100? it seems more suitable for AI-102.
upvoted 2 times
...
ranjsi01
2 years, 9 months ago
any easy way to understand this ?
upvoted 2 times
ning
2 years, 4 months ago
The images and videos will have varying sizes and formats. Normalization mean put them into the same dimension and same format images / videos before further processing
upvoted 1 times
...
...
spaceykacey
2 years, 12 months ago
"Local penalty detection models must be written by using BrainScript." BrainScript is used in Microsoft Cognitive Toolkit (CNTK) and it's network definition only supports batch normalization. So C is correct. https://docs.microsoft.com/en-us/cognitive-toolkit/batchnormalization
upvoted 3 times
...
gaint
3 years, 4 months ago
Not able to follow question and answer. This question will take atleast 15 mins to read and summarize :-).
upvoted 5 times
...
satishgunjal
3 years, 10 months ago
So Batch normalization is usefull to speedup the process where as normalization is use full to handle varying sizes and formats of input data.
upvoted 1 times
...
dmadhup
4 years, 6 months ago
Answer: C
upvoted 1 times
...
Community vote distribution
A (35%)
C (25%)
B (20%)
Other
Most Voted
A voting comment increases the vote count for the chosen answer by one.

Upvoting a comment with a selected answer will also increase the vote count towards that answer by one. So if you see a comment that you already agree with, you can upvote it instead of posting a new comment.

SaveCancel
Loading ...
exam
Someone Bought Contributor Access for:
SY0-701
London, 1 minute ago