Microsoft Azure AI-900 - AI Fundamentals Cards

Which two capabilities are supported natively by Azure Bot Service?

A bot can be used to respond to new student queries or to respond to questions via communication channels, such as email

What is Azure Bot Service?

A cloud-based platform for developing and managing conversational bots

You need to create an automated machine learning (automated ML) model. Which resource should you create first in Azure Machine Learning studio?

A dataset is required to create an automated machine learning (automated ML) run. A workspace must be created before you can access Machine Learning studio.

Vision Resource: Cognitive Services:

A general cognitive services resource that includes Computer Vision along with many other cognitive services, such as Text Analytics, Translator Text, and others. Use this resource type if you plan to use multiple cognitive services and want to simplify administration and development.

You need to use Azure Machine Learning to train a regression model. What should you create in Machine Learning studio?

A job must be created in Machine Learning studio to use Machine Learning to train a regression model. A workspace must be created before you can access Machine Learning studio.

Sentiment Analysis:

A neutral document would score around 0.5.

What is Azure Machine Learning?

A platform for training, deploying, and managing machine learning models.

You need to identify numerical values that represent the probability of humans developing diabetes based on age and body fat percentage. Which type of machine learning model should you use?

A possible type of machine learning model that you can use to predict the probability of humans developing diabetes based on age and body fat percentage is logistic regression.

Relative Absolute Error (RAE):

A relative metric between 0 and 1 based on the absolute differences between predicted and true values. The closer to 0 this metric is, the better the model is performing. Like RSE, this metric can be used to compare models where the labels are in different units.

Relative Squared Error (RSE):

A relative metric between 0 and 1 based on the square of the differences between predicted and true values. The closer to 0 this metric is, the better the model is performing. Because this metric is relative, it can be used to compare models where the labels are in different units.

Vision Resource: Computer Vision:

A specific resource for the Computer Vision service. Use this resource type if you don't intend to use any other cognitive services, or if you want to track utilization and costs for your Computer Vision resource separately.

What are Azure Cognitive Services?

A suite of services with four main pillars: Vision, Speech, Language, Decision

AUC = Area Under the Curve Scale:

AUC = Area Under the Curve Scale: 0.5: As good as random choice 0.5-0.7: Poor performance 0.7-0.8: OK performance >0.8-0.9: Very good performance

Common AI Workloads: Natural Language Processing:

Applications that can interpret written or spoken language, and engage in dialogs with human users

Common AI Workloads: Computer Vision:

Applications that interpret visual input from cameras, images, or videos

To create K clusters, you must set the number of ___________________to K

Centroids

Category Predictions =

Classification

Which three [3] data transformation modules are in the Azure Machine Learning designer?

Clean Missing Data, Normalize Data, Select Columns in a Dataset

Clustering machine learning models are used in many industries. A few scenarios are:

Cluster customer attribute data into segments for marketing analysis Cluster geographic coordinates into regions of high traffic in a city for a ride-share application Cluster written feedback into topics to prioritize customer service changes.

Clustering is unsupervised machine learning and automated ML only works with ___________________________learning algorithms.

Clustering

Clustering:

Clustering, in machine learning, is a method of grouping data points into similar clusters. It is also called segmentation. For example, you might apply clustering to find similar people by demographics. You might use clustering with text analysis to group sentences with similar topics or sentiment.

Microsoft Azure provides multiple cognitive services that you can use to detect and analyze faces, including:

Computer Vision, which offers face detection and some basic face analysis, such as returning the bounding box coordinates around an image. Video Indexer, which you can use to detect and identify faces in a video. Face, which offers pre-built algorithms that can detect, recognize, and analyze faces.

Some potential uses for computer vision include:

Content Organization: Identify people or objects in photos and organize them based on that identification. Text Extraction: Analyze images and PDF documents that contain text and extract the text into a structured format. Spatial Analysis: Identify people or objects, such as cars, in a space and map their movement within that space.

You need to use the Azure Machine Learning designer to deploy a predictive service from a newly trained model. What should you do first in the Machine Learning designer?

Create an inference pipeline.

What is Azure Cognitive Search?

Data extraction, enrichment, and indexing for intelligent search and knowledge mining

Which three [3] features are elements of the Language Service?

Entity Linking, PII detection, and sentiment analysis are all elements of the Azure Cognitive Service for Language.

Common AI Workloads: Knowledge Mining:

Extract information from data sources to create a searchable knowledge store.

Image Analysis Service:

Extracts many visual features from images, such as objects, faces, adult content, and auto generated text descriptions.

Additional Features are generated with:

Feature Engineering

Creating an object detection solution with Custom Vision consists of [3] three main tasks.

First you must use upload and tag images, then you can train the model, and finally you must publish the model so that client applications can use it to generate predictions.

Machine Learning:

IGNORE ME

Face Service:

If you need basic face detection and analysis, combined with general image analysis capabilities, you can use the Computer Vision service; but for more comprehensive facial analysis and facial recognition functionality, use the Face service.

Your company is in business of providing customer support over the phone. During many years company has created a frequently asked questions (FAQ) document. Company now wants to create a knowledge base that includes the questions and answers from the FAQ with the least possible effort. What should you do?

Import from the existing FAQ document into a new knowledge base.

Regression:

In the most basic sense, regression refers to prediction of a numeric target. Linear regression attempts to establish a linear relationship between one or more independent variables and a numeric outcome, or dependent variable. You use this module to define a linear regression method, and then train a model using a labeled dataset. The trained model can then be used to make predictions.

Name [3] Three guiding principles for responsible AI:

Inclusiveness, fairness, reliability & safety

Which two [2] features of Azure Cognitive Services allow you to identify issues from support question data, as well as identify any people and products that are mentioned?

Key phrase extraction is used to extract key phrases to identify the main concepts in a text. It enables a company to identify the main talking points from the support question data and allows them to identify common issues. Named entity recognition can identify and categorize entities in unstructured text, such as people, places, organizations, and quantities.

Which three [3] features are elements of the Speech service?

Language identification, speaker recognition, and voice assistants are all elements of the Speech service.

You need to identify numerical values that represent the probability of humans developing diabetes based on age and body fat percentage. Which type of machine learning model should you use?

Logistic regression which is a type of classification model, which returns either a Boolean value or a categorical decision.

Which feature of the Translator service is available only to Custom Translator?

Model training with a dictionary can be used with Custom Translator when you do not have enough parallel sentences to meet the 10,000 minimum requirements. The resulting model will typically complete training much faster than with full training and will use the baseline models for translation along with the dictionaries you have added.

[Y/N]: Automated machine learning can automatically infer the training data from the use case provided. -

NO

Uses of OCR: The ability to recognize printed and handwritten text in images, is beneficial in many scenarios such as:

Note taking, digitizing forms, such as medical records or historical documents, scanning printed or handwritten checks for bank deposits.

Which two [2] artificial intelligence (AI) workload features are part of the computer vision service? Each correct answer presents a complete solution.

OCR and Spatial Analysis are part of the computer vision service.

Bounding Box Coordinates =

Object Detection

The results from the Read API are arranged into the following hierarchy:

Pages - One for each page of text, including information about the page size and orientation Lines - The lines of text on a page, Words - The words in a line of text, including the bounding box coordinates and text itself.

When creating custom vision object detection models in Cognitive Services, At the end of the training process, the performance for the trained model is indicated by the following evaluation metrics:

Precision: What percentage of class predictions did the model correctly identify? For example, if the model predicted that 10 images are oranges, of which eight were actually oranges, then the precision is 0.8 (80%) Recall: What percentage of the class predictions made by the model were correct? For example, if there are 10 images of apples, and the model found 7 of them, then the recall is 0.7 (70%) Mean Average Precision (mAP): An overall metric that takes into account both precision and recall across all classes.

Use Classification models to:

Predict a yes or no answer is classification (is it yes or is it no?)

Common AI Workloads: Machine Learning:

Predictive models based on data and statistics — the foundation for Al

Numeric Value Predictions =

Regression

Which methods would you use to evaluate the performance of a model that has different unit!

Relative Squared Error (RSE): A relative metric between 0 and 1 based on the square of the differences between predicted and true values. The closer to 0 this metric is, the better the model is performing. Because this metric is relative, it can be used to compare models where the labels are in different units. Relative Absolute Error (RAE): A relative metric between 0 and 1 based on the absolute differences between predicted and true values. The closer to 0 this metric is, the better the model is performing. Like RSE, this metric can be used to compare models where the labels are in different units.

You have to learn the Confusion Matrix predicted class item in lower right.

Results Shown in Confusion Matrix - Memorize the diagram True Positive = 11 False Negative = 1033 False Positive = 5 True Negative = 13951 WATCH THIS VIDEO TO HELP YOU LEARN ABOUT CONFUSION MATRIX IN SEVEN MINS https://youtu.be/Kdsp6soqA7o

Sentiment Analysis:

Score values closer to 1 indicated a more positive sentiment where scores closer to 0 indicated negative sentiment.

Which type of artificial intelligence (AI) workload provides the ability to classify individual pixels in an image depending on the object that they represent?

Semantic segmentation provides the ability to classify individual pixels in an image depending on the object that they represent.

Because the Read API can work with large documents, it works asynchronously so as not to block your application while it is reading the content and returning results to your application. This means that to use the Read API, your application must use a [3] three-step process:

Submit an image to the API, and retrieve an operation ID in response. Use the operation ID to check on the status of the image analysis operation and wait until it has completed. Retrieve the results of the operation.

Common AI Workloads: Anomaly Detection:

Systems that detect unusual patterns or events, enabling pre-emptive action

Mean Absolute Error (MAE):

The average difference between predicted values and true values. This value is based on the same units as the label, in this case dollars. The lower this value is, the better the model is predicting.

Features & Labels: What are Features and Labels?

The features are descriptive attributes (serving as the input), while the label is the characteristic you are trying to predict (serving as the output).

Label vs. Features:

The output you get from your model after training is the LABEL but the FEATURES are patterns, colors, or forms that are a part of your class (feathers, fur etc.)

Root Mean Squared Error (RMSE):

The square root of the mean squared difference between predicted and true values. The result is a metric based on the same unit as the label (dollars). When compared to the MAE (above), a larger difference indicates greater variance in the individual errors (for example, with some errors being very small, while others are large).

Coefficient of Determination (R2):

This metric is more commonly referred to as R-Squared, and summarizes how much of the variance between predicted and true values is explained by the model. The closer to 1 this value is, the better the model is performing.

Which three supervised machine learning models can you train by using automated machine learning (automated ML) in the Azure Machine Learning studio?

Time-series forecasting, regression, and classification are supervised machine learning models.

Which three [3] supervised machine learning models can you train by using automated machine learning (automated ML) in the Azure Machine Learning studio?

Time-series forecasting, regression, and classification are supervised machine learning models. - Clustering is an unsupervised task.

You need to make the press releases of your company available in a range of languages: Which service should you use?

Translator Text

Classification:

Two-class classification provides the answer to simple two-choice questions such Classification as Yes/No or True/False. (This question is asked about a bank loan and a diabetes model, but it asks it both about a classification and a regression model, so pay attention to what it's actually asking for. Is it a value or a choice between two things or is it a numeric value?)

Which natural language processing (NLP) technique assigns values to words such as plant and flower, so that they are considered closer to each other than a word such as airplane?

Vectorization captures semantic relationships between words by assigning them to locations in n-dimensional space. Lemmatization, also known as stemming, normalizes words before counting them.

AUC = Area Under the Curve:

When using Azure ML designer pipeline to train and test a binary classification model, an AUC score of 0.3. means that the model performs worse than random guessing

Vectorization:

While Vectorization captures semantic relationships between words by assigning them to locations in n-dimensional space. Lemmatization, also known as stemming, normalizes words before counting them.

[Y/N]: Automated machine learning is the process of automating the time- consuming, iterative tasks of machine learning model development -

YES

[Y/N]: Automated machine learning works by running multiple training iterations that are scored and ranked by the metrics you specify. -

YES

Evaluate model is:

a component used to measure the accuracy of trained models.

Select Columns in Dataset is:

a data transformation component that is used to choose a subset of columns of interest from a dataset.

Split Data Module:

always use this to randomly split dataset into test and validation subsets

K-means clustering is:

an unsupervised machine learning algorithm component used for training clustering models. You can use unlabeled data with this algorithm.

Entities and intents:

are core components of a Language Understanding app model, but they are not used for testing the model.

Utterances:

are used to train and test a Language Understanding app model.

Linear regression:

attempts to establish a linear relationship between one or more independent variables and a numeric outcome, or dependent variable. You use this module to define a linear regression method, and then train a model using a labeled dataset. The trained model can then be used to make predictions.

Automated ML learning can predict categories or classes by using a _____________________ algorithm, as well as numeric values as part of the regression algorithm, and at a future point in time by using time-series data.

classification

The Text Analytics API is a __________________________service that provides advanced natural language processing over raw text, and includes four main functions: sentiment analysis, key phrase extraction, named entity recognition, and language detection.

cloud-based

An Azure container instance and an AKS cluster can be created as a deployment target, after training of a model is ___________________________.

complete.

Frequency analysis:

counts how often a word appears in a text. N-grams extend frequency analysis to include multi-term phrases.

The acoustic, language, and pronunciation scenarios require what?

developing your own model.

Optical Character Recognition:

extracts text from images and uses the read API to extract printed and handwritten text from documents and photos.

Hierarchical clustering:

groups data points that have similar characteristics.

What does Entity detection do?

identifies specific types of entity in the document, not the I main talking points.

Key phrases can be used to?

identify the main talking points in a text document.

Object detection vs. image classification

image classification is a machine learning based form of computer vision in which a model is trained to categorize images based on the primary subject matter they contain. Object detection goes further than this to classify individual objects within the image, and to return the coordinates of a bounding box that indicates the object's location.

Tagging:

involves associating an image with metadata that summarizes the attributes of the image. Detecting image types involves identifying clip art images or line drawings.

Azure Machine Learning Designer:

is a drag and drop interface to train and deploy models in Azure Machine Learning and it lets you add and connect modules on a visual canvas.

Logistic regression:

is a type of classification model, which returns either a Boolean value or a categorical decision.

Entity Linking:

is part of the entity recognition service, which returns links to external websites to disambiguate terms (entities) identified in a text.

Object Detection

is similar to tagging, but the API returns the bounding box coordinates (in pixels) for each object found in the image

Inclusiveness

keeps from unintentionally excluding people.

In a regression machine learning algorithm, what are the characteristics of features and labels in a training dataset?

known feature and label values.

Lemmatization, also known as stemming:

normalizes words before counting them. Frequency analysis counts how often a word appears in a text. N-grams extend frequency analysis to include multi-term phrases.

The Universal Language Model used by the speech-to-text API is:

optimized for conversational and dictation scenarios.

The Clean Missing Data module is:

part of preparing the data and data transformation process.

Sentiment analysis does what?

returns a numeric score indicating how positive or negative the text is.

Linear regression uses a _________________feature.

single

Form Recognizer:

uses advanced machine learning to accurately extract text, key value pairs, tables and structures from a document

Custom Vision:

uses your own images, image recognition service that allows you to improve your own image identifier