- (Exam Topic 1)
You need to implement a scaling strategy for the local penalty detection data. Which normalization type should you use?
Correct 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. References:
https://docs.microsoft.com/en-us/cognitive-toolkit/post-batch-normalization-statistics
- (Exam Topic 3)
You plan to create a speech recognition deep learning model. The model must support the latest version of Python.
You need to recommend a deep learning framework for speech recognition to include in the Data Science Virtual Machine (DSVM).
What should you recommend?
Correct Answer:
B
TensorFlow is an open source library for numerical computation and large-scale machine learning. It uses Python to provide a convenient front-end API for building applications with the framework
TensorFlow can train and run deep neural networks for handwritten digit classification, image recognition, word embeddings, recurrent neural networks, sequence-to-sequence models for machine translation, natural language processing, and PDE (partial differential equation) based simulations.
References:
https://www.infoworld.com/article/3278008/what-is-tensorflow-the-machine-learning-library-explained.html
- (Exam Topic 1)
You need to implement a feature engineering strategy for the crowd sentiment local models. What should you do?
Correct Answer:
D
The linear discriminant analysis method works only on continuous variables, not categorical or ordinal variables.
Linear discriminant analysis is similar to analysis of variance (ANOVA) in that it works by comparing the means of the variables.
Scenario:
Data scientists must build notebooks in a local environment using automatic feature engineering and model building in machine learning pipelines.
Experiments for local crowd sentiment models must combine local penalty detection data. All shared features for local models are continuous variables.
- (Exam Topic 3)
You are performing sentiment analysis using a CSV file that includes 12,000 customer reviews written in a short sentence format. You add the CSV file to Azure Machine Learning Studio and configure it as the starting point dataset of an experiment. You add the Extract N-Gram Features from Text module to the experiment to extract key phrases from the customer review column in the dataset.
You must create a new n-gram dictionary from the customer review text and set the maximum n-gram size to trigrams.
What should you select? To answer, select the appropriate options in the answer area.
NOTE: Each correct selection is worth one point.
Solution:
Vocabulary mode: Create
For Vocabulary mode, select Create to indicate that you are creating a new list of n-gram features. N-Grams size: 3
For N-Grams size, type a number that indicates the maximum size of the n-grams to extract and store. For example, if you type 3, unigrams, bigrams, and trigrams will be created.
Weighting function: Leave blank
The option, Weighting function, is required only if you merge or update vocabularies. It specifies how terms in the two vocabularies and their scores should be weighted against each other.
References:
https://docs.microsoft.com/en-us/azure/machine-learning/studio-module-reference/extract-n-gram-features-from
Does this meet the goal?
Correct Answer:
A
- (Exam Topic 2)
You need to configure the Feature Based Feature Selection module based on the experiment requirements and datasets.
How should you configure the module properties? To answer, select the appropriate options in the dialog box in the answer area.
NOTE: Each correct selection is worth one point.
Solution:
Box 1: Mutual Information.
The mutual information score is particularly useful in feature selection because it maximizes the mutual information between the joint distribution and target variables in datasets with many dimensions.
Box 2: MedianValue
MedianValue is the feature column, , it is the predictor of the dataset.
Scenario: The MedianValue and AvgRoomsinHouse columns both hold data in numeric format. You need to select a feature selection algorithm to analyze the relationship between the two columns in more detail.
References:
https://docs.microsoft.com/en-us/azure/machine-learning/studio-module-reference/filter-based-feature-selection
Does this meet the goal?
Correct Answer:
A
