AI FUNDAMENTALS Flashcards

(61 cards)

1
Q

Which types of tasks AI CAN DO

A

Tasks which requires human intelligence like
Visual perception
Speech recognition
Decision making
Translation between languages

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
2
Q

Which Type of ai focus on specific task

A

Narrow AI or weak AI (E.G. VIRTUAL ASSISTANT, self driving cars ,Machine learning (learn from data (examples)

Wheras STRONG AI or genral AI in which
Intelligence of machine = intelligence of human
Like the machine that can solve problems,learn and plan for future

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
3
Q

Is machine learning focus on rules ?

A

No, machine learning is learning from examples( using those examples ,creating model and use that model to make predictions)or use historical data to predict future

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
4
Q

Name challenges faced in machine learning

A
  1. No. of examples needed
  2. Availability of skilled personnel
  3. Complexity in implementing MLOPS
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
5
Q

What are the 3 approaches to building AI solutions in AZURE

A
  1. Use pre trained models (get intelligence from text,images,audio,video)
    I.e azure ai services (language,vision,face, speech)
  2. Build simple models (without needing data scientist) :no or limited code experience
    Azure ai vision(custom vision)
    Azure ML OR Automated ml
  3. Build complex models (using data scientist)
    Using azure ML : build your own ml models from zero
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
6
Q

Which service used to get intelligence from images

A

Azure AI VISION SERVICE and it had different api:
1. Analyze Image API
2. Read OCR API
3. Custom vision

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
7
Q

Which nlp workload is used to perform better on small amounts of text:
1. Key phrase extraction
2. Sentiment analysis
3. Speech synthesis

A

Ans: b

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
8
Q

Which is used to interpret written text application

A

Nlp

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
9
Q

What type of nlp solution is demonstrated by incorporating synthesized neural voice into your robotic system

A

Text to speech

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
10
Q

What is an example of assessing vehicle damage from a photograph

A

Anomalies detection

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
11
Q

Which service used to extract data from handwritten letters

A

Ocr

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
12
Q

Which scenarios are governed by accountability principle in responsible AI

A
  1. Meet ethical standards
    2.Following a governance network
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
13
Q

Which responsible AI principle focusses on reducing bias in ml model

A

Fairness

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
14
Q

Which AI principle mandates rigorous testing of an ai based application to mitigate potential risks to human life

A

Reliability and safety

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
15
Q

Which 2 guiding principles are included in Microsoft’s approach to responsible AI

A

Inclusiveness
Reliability and safety

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
16
Q

Which option describe a ml method used for grouping data points that share similar traits

A

Clustering

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
17
Q

How do you assess performance of a classification model

A

True positive rate

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
18
Q

Which 2 components are draggable onto a canvas within azure machine learning designer

A

Dataset
Module

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
19
Q

_ is ensuring that an AI SYSTEM does not provide a prediction when important fields contain unusual or missing values a principle for responsible AI

A

Reliability n safety

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
20
Q
  • is ensuring that the numeric variables in training data are on a similar scale and example of what
A

Feature engineering

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
21
Q

_ use to assigning classes to images before training a classification model

A

Labelling

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
22
Q

In which model ,label must be numeric
1. Regression
2.Classification
3. Clustering

A

Ans: Regression

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
23
Q

Which model can be used to forecast the sale price of auctioned items

A

Regression

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
24
Q

Which service should you use to determine a user’s intent

A

AI Language

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
25
Which type of computer vision used to identify street signs in photograph
Optical character recognition
26
Detecting unusual temprature fluctuations for a large machine is an example of _
Anomaly detection workload
27
Which type of NLP workload that analyze social media post to extract mention of city names and city names discussed frequently
Entity recognition
28
In Ml model,data that is used as inputs are _
Variables
29
Ensuring an AI SYSTem does not provide a prediction when important fields contain unusual or missing values is _ principle of responsible AI
Reliability and safety
30
You need to extract text from the images ,which azure service will you be using
Azure AI VISION
31
Returning a bounding box that indicates location of vehicle in an image is an example of _
Object detection
32
Which AI SERVICE is unified service for investing and transforming data of diffrent modalities ,business can extract insights from documents,images ,videos n audio seamlessly
Azure AI Content Understanding
33
_ is unified platform designed for developing and deploying gen ai application and azure ai api's
Azure AI foundry
34
_ metric used in regression to measure the average magnitude of errors between predicted and actual values
Means absolute error(MAE) ( Once task shifted from regression to classification then MAE is no LONGER applicable)
35
Which service used to detect and filter harmful content
Azure AI Content safety
36
Which service used to monitor the metrics like time series data and detect anomalies in data streams
Azure AI metrics advisor
37
Which service used to do image recognition and analysis tools and extract insights from visual data
Azure AI vision
38
Which services contains pretrained generative models like GPT AND can fine tune and integrate into our applications
Azure OpenAI
39
Which nlp technique normalised words before counting them
Stemming
40
Which 2 ai workload features are prt of azure vision in foundry tool service
Optical character recognition Spatial analysis
41
Which type of workload focus on searching and indexing of data
Data mining
42
Which principle of responsible AI raise awareness about the limitations of AI BASED SOLUTIONS
Transparency
43
Which feature of azure language in foundry tool service includes functionality that return links to external websites to disambiguate terms identified in a text
Entity recognition
44
Which 3 values returned by language detection feature of azure language in foundry tool service in azure
ISO 6391 CODE Language Name Score
45
Which 3 features are elements of azure language in foundry tool service
Entity linking Personal identifiable information detection Sentiment analysis
46
Which 3 features are elements of azure speech in foundry tool service
Language identification Speaker recognition Voice assistants
47
What allows you to identify different vehicle types in traffic monitoring images
Object detection
48
_ service used to eliminate the need for choosing,training and evaluating a model by providing pre-trained models
Computer vision (i.e. Azure AI vision)
49
Which analytical task of azure vision in foundry tool service returns bounding box co-ordinates
Object detection
50
Which feature of computer vision involves associating an image with metadata that summarised attribute of image
Tagging
51
You need to use the azure machine learning designer to deploy a predictive service from a newly trained models. What should you do first in machine learning designer
Create an inference pipeline
52
What are the examples of supervised machine learning models
Classification Regression Time series forecasting
53
_ can used to identify constraint and style for the responses of a gen ai model
System messages
54
_ can search ,classify and compare sources of text for similarity
Embeddings
55
Which preprocessing step would most improve OCR Accuracy for photos taken in low light conditions
Applying histogram equalization
56
An AI SYSTem for autonomous vehicle decision making needs to balance multiple responsible AI dimensions. Which sequence of priorities would create most appropriate balance for this application
Safety - < Controllability-< Transparency
57
In responsible AI development, achieving _ required implementing processes that define and enforce practices while ensuring models remain beneficial for both individuals and society
Governance and safety
58
A research organisation needs to implement semantic search across multiple languages for their scientific documents.Which sequence of knowledge mining capabilities would BEST address this requirement
OCR - Translation - Entity Recognition -Indexing
59
_ is an AI FRAMEWORK that involves retrieval of relevant information from external sources to inform and enhance the generation of responses
Retrieval Augmented generation(RAG)
60
For which 2 scenarios is the Universal Language Model used by speech-to-text API optimised.
Conversational Dictation
61
Which principle of responsible AI plays the primary role when implementing an AI solution rfqr meet qualification of business loan approvals
Faieness