Machine Learning Engineering Associate 2

Question 1

Data Wrangler

Answer 1

Visual data preparation tool in Amazon SageMaker for exploring; transforming; and analyzing data

Question 2

Glue

Answer 2

Fully managed extract; transform; and load (ETL) service

Question 3

Glue DataBrew

Answer 3

Visual data preparation tool that makes it easy to clean and normalize data

Question 4

Kinesis

Answer 4

Platform for streaming data on AWS

Question 5

Lambda

Answer 5

Serverless compute service for running code without provisioning servers

Question 6

SageMaker Ground Truth

Answer 6

Fully managed data labeling service for building accurate training datasets

Question 7

Class imbalance

Answer 7

Situation where classes in a dataset are not represented equally

Question 8

Server-side encryption

Answer 8

Data encryption performed by the storage service

Question 9

Client-side encryption

Answer 9

Data encryption performed by the client before sending to storage

Question 10

Data anonymization

Answer 10

Removing or encrypting personally identifiable information from datasets

Question 11

Supervised learning

Answer 11

ML approach where the model is trained on labeled data

Question 12

Unsupervised learning

Answer 12

ML approach where the model is trained on unlabeled data

Question 13

Reinforcement learning

Answer 13

ML approach where an agent learns to make decisions by interacting with an environment

Question 14

Feature importance

Answer 14

Measure of how much each feature contributes to the model’s predictions

Question 15

SHAP values

Answer 15

Shapley Additive exPlanations; a game theoretic approach to explain machine learning model outputs

Question 16

XGBoost

Answer 16

Gradient boosting algorithm known for speed and performance

Question 17

Epoch

Answer 17

One complete pass through the entire training dataset

Question 18

Early stopping

Answer 18

Technique to stop training when performance on a validation set stops improving

Question 19

Distributed training

Answer 19

Spreading the training process across multiple compute resources

Question 20

Hyperparameter tuning

Answer 20

Process of finding the best combination of hyperparameters for a model

Question 21

Transfer learning

Answer 21

Using knowledge gained from solving one problem to solve a related problem

Question 22

Dropout

Answer 22

Technique where randomly selected neurons are ignored during training

Question 23

Weight decay

Answer 23

Adding a penalty term to the loss function to prevent overfitting

Question 24

Random search

Answer 24

Randomly sampling hyperparameters from a defined search space

Question 25

Bayesian optimization

Answer 25

Using probabilistic model to guide the search for optimal hyperparameters

Question 26

Confusion matrix

Answer 26

Table showing correct and incorrect predictions for each class

Question 27

F1 score

Answer 27

Harmonic mean of precision and recall

Question 28

ROC

Answer 28

Graph showing the performance of a classification model at all classification thresholds

Question 29

AUC

Answer 29

Measure of the ability of a classifier to distinguish between classes

Question 30

Overfitting

Answer 30

Model performs well on training data but poorly on unseen data

Question 31

Underfitting

Answer 31

Model performs poorly on both training and unseen data

Question 32

Concept drift

Answer 32

Changes in the underlying relationships between input and output variables

Question 33

Data drift

Answer 33

Changes in the statistical properties of the input data

Question 34

A/B testing

Answer 34

Experiment where two variants of a model are compared to determine which performs better

Question 35

CloudTrail

Answer 35

Service that records API calls and other account activity in AWS

Question 36

Cost Explorer

Answer 36

Tool for visualizing; understanding; and managing AWS costs and usage over time

Question 37

IAM roles

Answer 37

Set of permissions that define what actions are allowed or denied in AWS

Question 38

Security groups

Answer 38

Virtual firewalls for controlling inbound and outbound traffic to AWS resources

Question 39

Network ACLs

Answer 39

Optional layer of security that acts as a firewall for controlling traffic in and out of subnets

Question 40

Least privilege access

Answer 40

Principle of giving users the minimum levels of access necessary to complete their tasks