Basics

Question 1

Machine learning

Answer 1

his is often the foundation for an AI system, and is the way we “teach” a computer model to make predictions and draw conclusions from data.

Question 2

Computer vision

Answer 2

Capabilities within AI to interpret the world visually through cameras, video, and images.

Question 3

Natural language processing

Answer 3

Capabilities within AI for a computer to interpret written or spoken language, and respond in kind.

Question 4

Document intelligence

Answer 4

Capabilities within AI that deal with managing, processing, and using high volumes of data found in forms and documents.

Question 5

Knowledge mining

Answer 5

Capabilities within AI to extract information from large volumes of often unstructured data to create a searchable knowledge store.

Question 6

Generative AI

Answer 6

Capabilities within AI that create original content in a variety of formats including natural language, image, code, and more.

Question 7

Automated machine learning:

Answer 7

this feature enables non-experts to quickly create an effective machine learning model from data.

Question 8

Azure Machine Learning designer

Answer 8

a graphical interface enabling no-code development of machine learning solutions.

Question 9

Semantic segmentation

Answer 9

Semantic segmentation is an advanced machine learning technique in which individual pixels in the image are classified according to the object to which they belong. For example, a traffic monitoring solution might overlay traffic images with “mask” layers to highlight different vehicles using specific colors.

Question 10

Data metric visualization:

Answer 10

analyze and optimize your experiments with visualization.

Question 11

Notebooks

Answer 11

write and run your own code in managed Jupyter Notebook servers that are directly integrated in the studio.

Question 12

how is document intelligence used

Answer 12

reading and processing documents like filling out forms, checking forms, finding the right documents in massive scanned document lists

Question 13

Image analysis

Answer 13

You can create solutions that combine machine learning models with advanced image analysis techniques to extract information from images, including “tags” that could help catalog the image or even descriptive captions that summarize the scene shown in the image.

Question 14

How is knowledge mining used?

Answer 14

creating indexes of large number of internal/ external documents to improve search, nlp and tagging for finding documents

Question 15

Supervised machine learning

Answer 15

Supervised machine learning is a general term for machine learning algorithms in which the training data includes both feature values and known label values. Supervised machine learning is used to train models by determining a relationship between the features and labels in past observations, so that unknown labels can be predicted for features in future cases

Question 16

Regression

Answer 16

Regression is a form of supervised machine learning in which the label predicted by the model is a numeric value. For example:

Question 17

Classification

Answer 17

Classification is a form of supervised machine learning in which the label represents a categorization, or class. There are two common classification scenarios.

Question 18

Binary classification

Answer 18

n binary classification, the label determines whether the observed item is (or isn’t) an instance of a specific class. Or put another way, binary classification models predict one of two mutually exclusive outcomes.

Question 19

Multiclass classification

Answer 19

Multiclass classification extends binary classification to predict a label that represents one of multiple possible classes.

Question 20

Unsupervised machine learning

Answer 20

Unsupervised machine learning involves training models using data that consists only of feature values without any known labels. Unsupervised machine learning algorithms determine relationships between the features of the observations in the training data.

Question 21

Clustering

Answer 21

The most common form of unsupervised machine learning is clustering. A clustering algorithm identifies similarities between observations based on their features, and groups them into discrete clusters

Question 22

what are the 4 key elements in a training of a supervised model

Answer 22

Split the data into training and testing
Use an algorithm to fit the data to the model
use the validation to testing the model against the subset
Compare results

Question 23

What is Mean Absolute Error (MAE)

Answer 23

The variance in this example indicates by how many ice creams each prediction was wrong. It doesn’t matter if the prediction was over or under the actual value (so for example, -3 and +3 both indicate a variance of 3). This metric is known as the absolute error for each prediction, and can be summarized for the whole validation set as the mean absolute error (MAE).

It is an average of the error between predicted and actuall results

Question 24

Mean Squared Error (MSE)

Answer 24

The mean absolute error metric takes all discrepancies between predicted and actual labels into account equally. However, it may be more desirable to have a model that is consistently wrong by a small amount than one that makes fewer, but larger errors. One way to produce a metric that “amplifies” larger errors by squaring the individual errors and calculating the mean of the squared values. This metric is known as the mean squared error (MSE).

Question 25

Root Mean Squared Error (RMSE)

Answer 25

The mean squared error helps take the magnitude of errors into account, but because it squares the error values, the resulting metric no longer represents the quantity measured by the label. In other words, we can say that the MSE of our model is 6, but that doesn’t measure its accuracy in terms of the number of ice creams that were mispredicted; 6 is just a numeric score that indicates the level of error in the validation predictions.

Question 26

Coefficient of determination (R2)

Answer 26

All of the metrics so far compare the discrepancy between the predicted and actual values in order to evaluate the model. However, in reality, there’s some natural random variance in the daily sales of ice cream that the model takes into account. In a linear regression model, the training algorithm fits a straight line that minimizes the mean variance between the function and the known label values. The coefficient of determination (more commonly referred to as R2 or R-Squared) is a metric that measures the proportion of variance in the validation results that can be explained by the model, as opposed to some anomalous aspect of the validation data (for example, a day with a highly unusual number of ice creams sales because of a local festival).

Question 27

confusion matrix

Answer 27

a matrix to compare TP, TN, FP, FN
(true positive, True neg, etc)

Question 28

Accuracy

Answer 28

The simplest metric you can calculate from the confusion matrix is accuracy - the proportion of predictions that the model got right. Accuracy is calculated as:

(TN+TP) ÷ (TN+FN+FP+TP)

Question 29

Recall

Answer 29

call is a metric that measures the proportion of positive cases that the model identified correctly. In other words, compared to the number of patients who have diabetes, how many did the model predict to have diabetes?

The formula for recall is:

TP ÷ (TP+FN)

Question 30

Precision

Answer 30

Precision is a similar metric to recall, but measures the proportion of predicted positive cases where the true label is actually positive. In other words, what proportion of the patients predicted by the model to have diabetes actually have diabetes?

The formula for precision is:

TP ÷ (TP+FP)

Question 31

F1-score

Answer 31

F1-score is an overall metric that combined recall and precision. The formula for F1-score is:

(2 x Precision x Recall) ÷ (Precision + Recall)

Question 32

Area Under the Curve (AUC)

Answer 32

Another name for recall is the true positive rate (TPR), and there’s an equivalent metric called the false positive rate (FPR) that is calculated as FP÷(FP+TN).

Question 33

Deep Learning

Answer 33

A process that mimics the human mind, in a limited fashion. We train a model (or fit a model) so that inputs have weights that impact the observation. A higher weight has more impact.

Question 34

What is the loss function?

Answer 34

loss function determines the overall variance, or loss, between predicted and actual label values.

Question 35

What is a feature?

Answer 35

Feature is the model input - these are the fields in the data that drive an answer (a label)

Question 36

What is a label

Answer 36

A label is the answer of a model - a yes/no or a value from a list. We look at features to predict a label

Question 37

what is feature selection?

Answer 37

during preprocessing you determine features that influence the prediction.

Question 38

what algorithm is used for clustering

Answer 38

k-means clustering

Question 39

what are the 4 typical steps for model training

Answer 39

Feature selection
removing data outliers
impute missing values
normalize numeric features

Question 40

the principle that describes raising awareness of the limitations of responsible AI-based
solutions is called

Answer 40

transparency

Question 41

What capabilities does azure cognitive services support

Answer 41

Azure Text Analytics supports chatbot integration,
multilingual content, and
confidence scoring.
It recognizes about 120 languages. Document sizes must
be under 5,120 characters.

Question 42

What is chitchat?

Answer 42

You can add personality to a chatbot by providing answers that use a specific
conversational tone. You use the chitchat feature to add the answers to a
chatbot knowledge base.

Question 43

Which is a multiclass classification algorithm

Answer 43

Decision Forest

Question 44

Principal Component Analysis is

Answer 44

to simplify complex data sets by finding the most important features or dimensions.

Question 45

The fraction of time when the model is correct is known as:

Answer 45

Accuracy

Question 46

Which of these confirms how often the model is correct:

Answer 46

Precision

Question 47

Which value identifies how much the model finds all there is to find?

Answer 47

Recall