AWS AI Practitioner Flashcards

1
Q

Where can I get free AWS AI Practitioner Certification training?

A

AI Practitioner Practice Exam Questions

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

What level is the AWS AI Practitioner certification?

A

Foundational

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

What are the AWS AI Practitioner Testing specifics?

A

120 minutes for 85 questions

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

5 Domains of AI Practitioner Exam

A

AI & ML (20%)
Generative AI (24%)
Applications of Foundation Models (28%)
Responsible AI (14%)
AI Security Compliance & Governance (14%)

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

Define AI

A

Any way a computer can mimic human intelligence

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
6
Q

Define ML

A

AI subset. Identifies patterns within data. Uses patterns to predict new patterns.

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

Define Deep Learning

A

ML subset. Neural Networks. Trying to replicate how the brain works.

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

Define Generative AI

A

DL subset. Uses DL to generate content.

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

How is traditional programming different than ML?

A

Traditional uses human to write the code.
ML uses a learning algorithm to create a model based on its learnings to produce a prediction.

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

When to use ML vs. Traditional programming

A
  1. When you cant code it (too complex to write the code)
  2. When you cant scale it (e.g. fraud detection, spam, recommendations)
  3. When you have to adapt/personalize (e.g. book recommendations)
  4. When you cant track it (data coming in to quickly, e.g. automated driving)
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
11
Q

Conditional ML is _______ in its predictions

A

Deterministic

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

Generative IA is ________ in its predictions

A

Non-Deterministic. Can make things up

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
13
Q

Why use ML over Gen AI?

A

Transparency * Interpretability
Explainability & Fairness
Robustness & Consistency
Data Efficiency
Specific Task Performance

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

3 Machine Learning Types

A

Supervised, Unsupervised, Reinforcement

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

What is ML/Supervised Learning

A

Data is assigned a label. Supervised learning uses the labels and repetition to learn. (E.g. how little kids learn).
Labeled data is put into a ML algorithm which makes a prediction. A human corrects /adjusts the model if the prediction is incorrect.

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

What is unsupervised learning?

A

No labeled data. Uses grouping data to predict other relationships. E.g. these type customers bought x so they will likely buy Y

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

What is ML/Reinforcement learning?

A

Learns through reinforcement. E.g. Dog training.

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

Key areas of ML/Supervised Learning

A

Regression (Numeric e.g. predict house prices), Classification (Binary (a or b class) or Multiclass (can fit into several classes)

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

Key area of ML/Unsupervised Learning

A

Grouping/Clustering (Labels unknown or Finds data in Patterns)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
20
Q

Key area of Reinforcement learning

A

Correct actions rewarded (if agent does x, do y)

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

ML/Self-Supervised Learning

A

The basis of Generative AI. Unlabled data has a word removed. The model then predicts the word that was removed. Do this over and over and the prediction becomes accurate.

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

Label/Target

A

Dependent variable: what you are attempting to predict

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

Feature

A

Independent variable: data that helps you make predictions

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

Feature Engineering

A

Data Transformation: process of reshaping data to get more value

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
25
Q

Feature Selection

A

Variable/subset selection: process of using the most valuable data

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
26
Q

ML Development Lifecycle

A

Problem/ML Problem Framing/Data Collection/Data Integration/Data Prep/Data viz & analysis

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
27
Q

Key Benefit of Sagemaker

A

It manages the entire ML lifecycle

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
28
Q

Rekognition

A

ML service: Image and video analysis

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
29
Q

Textract

A

ML service: extracts text and data from documents (e.g. OCR)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
30
Q

Comprehend

A

ML service: discovers insights and relationships in text

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
31
Q

Kendra

A

ML service: machine learning search service that allows you to use natural language questions

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
32
Q

Personalize

A

ML service: create personalized experiences (e.g. retail site - prompt new product discovery)

33
Q

Fraud Detector

A

ML service: helps id fraudulent activity

34
Q

What would you use to develop a learning model to analyze customer feedback on products. Customer will train the ml model

A

Supervised Learning

35
Q

Ecommerce company needs to integrate tailored recommendations to customers based on browsing and purchase history.

A

Amazon Personalize

36
Q

A ML engineer wants to implement a ml pipeline on AWS to automate training and deploying models. Pipeline should include data preprocessing, training and model deployment

A

Sagemaker

37
Q

Zero-Shot Prompting

A

The larger the LLM, the more likely the zero-shot prompt will yield effective results

38
Q

What improves zero-shot tuning

A

Instruction Tuning

39
Q

Few-Shot Prompting

A

Provide several examples to help the model

40
Q

Chain-of-thought prompting

A

Use CoT prompting when the task involves several steps or a requires a series of reasoning

41
Q

What is prompt injection?

A

When an attacker

42
Q

What is prompt leaking?

A

When we leak sensitive information

43
Q

What are two common malicious activities for AI

A

Prompt Injection and Prompt Leaking

44
Q

What is a prompt template?

A

Use when using prompts for large data sets when you are applying several variations of a prompt.

45
Q

RAG

A

Retrieval Augmented Generation`

46
Q

What is RAG?

A

Framework for building generative AI applications that can make use of data sources

47
Q

What is a Transformer?

A
48
Q

What is a Token?

A
49
Q

What is prompt engineering?

A
50
Q

What are the costs with the different prompt engineering options?

A
51
Q

What is the use case for Bedrock guardails?

A
52
Q

What makes up responsible AI?

A

Fairness, Explainability, Controllability, Safety, Privacy & Security, Governance, Transparency, Veracity & Robustness

53
Q

What is dataset Bias?

A

The imbalance in data used to train a ML model

54
Q

What are common types of dataset Bias?

A

Sampling, Historical, Measurement

55
Q

What is Sample Bias?

A

Data does not represent the true population

56
Q

What is historical Bias?

A

Data reflects past biases and inequities in society

57
Q

What is Measurement Bias?

A

the data collection process introduces errors

58
Q

I can you identify imbalances in data sets

A

Calculate the ration of the smaller class vs. total data

59
Q

What are model generalization problem examples

A

Underfitting, Overfitting, Appropriate Fitting

60
Q

What is Underfitting?

A

model is too simple

61
Q

How do you identify Underfitting?

A

Poor training and test results

62
Q

How do you fix Underfitting?

A

Increase Training Data or passes through the existing data

63
Q

What is Overfitting?

A

Model picks up noise instead of underlying relationships

64
Q

How do you identify Overfitting?

A

Model training does good. When you add more data, the model stops working.H

65
Q

How do you fix Overfitting?

A

Reduce model flexibility: fewer features, smaller groups of data per training job, adjust weights

66
Q

What type of bias and variance (H,M,L) is ideal for a model

A

Low bias and low variance

67
Q

AWS ML visually explained

A

https://mlu-explain.github.io/

68
Q

What is a linear regression?

A

What is

69
Q

What is Reinforcement Learning?

A

What

70
Q

what are the stages in governing the model lifecycle?

A

Onboard, Build, Train, Deploy, Monitor,

71
Q

What is ROC?

A
72
Q

What is A2I?

A

Amazon Augmented AI. Provides a human review of ML predictions

73
Q

What is SageMaker Clarify?

A

Helps to detect Bias in ML.

74
Q
A

ID data imbalances, check trained model for bias, explain overall behavior

75
Q

SageMaker Model Monitor

A

Monitors ML models in production. Detects drift and quality issues

76
Q

What is SageMaker Role Maker?

A

Define minimum permissions using premade permission sets.

77
Q

What are SageMaker Model Cards?

A

Document, retrieve and share model information

78
Q

What is SageMaker Model Dashboard?

A

Monitor model performance through a unified view.