MLA-C01

Question 1

Before you can use auto scaling, you must have already created an Amazon SageMaker ______________.

Answer 1

model endpoint.

Question 2

You can have multiple model _____________for the same endpoint.

Answer 2

versions

Question 3

Amazon SageMaker ____________ provides tools to help explain how machine learning (ML) models make predictions.

Answer 3

Clarify

Question 4

An ____________can be thought of as the answer to a Why question that helps humans understand the cause of a prediction.

Answer 4

explanation

Question 5

On AWS, AI/ML practitioners can use Amazon Sagemaker ____________, which uses Shapley values to help answer how different variables influence model behavior.

Answer 5

Clarify

Question 6

Debug model output tensors from machine learning training jobs in real time and detect non-converging issues using Amazon SageMaker ____________.

Answer 6

Debugger

Question 7

___________is the extent to which you can explain the internal mechanics of an ML or deep learning system in human terms.

Answer 7

Explainability

Question 8

Amazon SageMaker _________produces metrics that measure the predictive quality of machine learning model candidates.

Answer 8

Autopilot

Question 9

The ratio of the number of correctly classified items to the total number of (correctly and incorrectly) classified items.

Answer 9

Accuracy

Question 10

measures how well an algorithm predicts the true positives (TP) out of all of the positives that it identifies.

Answer 10

Precision

Question 11

uses natural language processing (NLP) to extract insights about the content of documents. It develops insights by recognizing the entities, key phrases, language, sentiments, and other common elements in a document.

Answer 11

Amazon Comprehend

Question 12

a text translation service that uses advanced machine learning technologies to provide high-quality translation on demand. use to translate unstructured text documents or to build applications that work in multiple languages.

Answer 12

Amazon Translate

Question 13

a fully managed, automatic speech recognition (ASR) service that makes it easy for developers to add speech to text capabilities to their applications.

Answer 13

Amazon Transcribe

Question 14

a cloud service that converts text into lifelike speech. You can use to develop applications that increase engagement and accessibility.

Answer 14

Amazon Polly

Question 15

a cloud-based image and video analysis service that makes it easy to add advanced computer vision capabilities to your applications.

Answer 15

Amazon Rekognition

Question 16

a fully managed service that uses statistical and machine learning algorithms to deliver highly accurate time-series forecasts.

Answer 16

Amazon Forecast

Question 17

an AWS service for building conversational interfaces for applications using voice and text.

Answer 17

Amazon Lex

Question 18

a fully managed machine learning service that uses your data to generate item recommendations for your users. It can also generate user segments based on the users’ affinity for certain items or item metadata.

Answer 18

Amazon Personalize

Question 19

a machine learning (ML) service that automatically extracts text, handwriting, and data from scanned documents.

Answer 19

Amazon Textract

Question 20

an intelligent search service that uses natural language processing and advanced machine learning algorithms to return specific answers to search questions from your data.

Answer 20

Amazon Kendra

Question 21

allows you to conduct a human review of machine learning (ML) systems to guarantee precision.

Answer 21

Amazon Augmented AI (A2I)

Question 22

uses machine learning (ML) to make it easier for customers to accurately detect anomalies in their metrics.

Answer 22

Amazon Lookout for Metrics

Question 23

a fully managed service enabling customers to identify potentially fraudulent activities and catch more online fraud faster.

Answer 23

Amazon Fraud Detector

Question 24

a fully managed, generative-AI powered assistant that you can configure to answer questions, provide summaries, generate content, and complete tasks based on your enterprise data.

Answer 24

Amazon Q Business

Question 25

Amazon Polly is the Opposite of Amazon ____________.

Answer 25

Transcribe

Question 26

______________measures how many actual positives were predicted as positive.

Answer 26

Recall

Question 27

_____________is the harmonic mean of precision and recall.

Answer 27

F1-measure

Question 28

It measures the ability of the model to predict a higher score for positive examples as compared to negative examples.

Answer 28

AUC (Area Under Curve)

Question 29

_________is a method used in machine learning to reduce errors in predictive data analysis.

Answer 29

Boosting

Question 30

____________improves machine models' predictive accuracy and performance by converting multiple weak learners into a single strong learning model.

Answer 30

Boosting

Question 31

____________ are data structures in machine learning that work by dividing the dataset into smaller and smaller subsets based on their features

Answer 31

Decision trees

Question 32

Boosting creates an ____________model by combining several weak decision trees sequentially.

Answer 32

ensemble

Question 33

In ________, data scientists improve the accuracy of weak learners by training several of them at once on multiple datasets. In contrast, boosting trains weak learners one after another.

Answer 33

bagging

Question 34

__________is a popular and efficient open-source implementation of the gradient boosted trees algorithm.

Answer 34

XGBoost

Question 35

___________boosting is a supervised learning algorithm that tries to accurately predict a target variable by combining multiple estimates from a set of simpler models.

Answer 35

Gradient

Question 36

Amazon SageMaker _____________ reduces data prep time for tabular, image, and text data from weeks to minutes.

Answer 36

Data Wrangler

Question 37

With SageMaker ________________ you can simplify data preparation and feature engineering through a visual and natural language interface.

Answer 37

Data Wrangler

Question 38

Sagemaker ____________ a no-code ML tool that helps business analysts generate accurate ML predictions without having to write code or without requiring any ML experience.

Answer 38

Canvas

Question 39

Amazon SageMaker ____________ is a fully managed, purpose-built repository to store, share, and manage features for machine learning (ML) models.

Answer 39

Feature Store

Question 40

___________are inputs to ML models used during training and inference.

Answer 40

Features

Question 41

SageMaker ____________ tags and indexes feature groups so they are easily discoverable through the visual interface of Amazon SageMaker Studio.

Answer 41

Feature Store

Question 42

Amazon SageMaker ____________ offers the most comprehensive set of human-in-the-loop capabilities, allowing you to harness the power of human feedback across the ML lifecycle to improve the accuracy and relevancy of models.

Answer 42

Ground Truth

Question 43

You can complete a variety of human-in-the-loop tasks with SageMaker ___________, from data generation and annotation to model review, customization, and evaluation, either through a self-service or an AWS-managed offering.

Answer 43

Ground Truth

Question 44

SageMaker _________helps identify potential bias during data preparation without writing code.

Answer 44

Clarify

Question 45

SQL function used for anomaly detection on numeric columns in a stream

Answer 45

RANDOM_CUT_FOREST

Question 46

is derived from “Linux” and “cluster

Answer 46

Lustre

Question 47

a type of parallel distributed file system, for large-scale computing

Answer 47

Lustre

Question 48

a fully managed Windows file system share drive

Answer 48

FSx for Windows File Server

Question 49

a network drive you can attach to your instances while they run

Answer 49

EBS

Question 50

Managed NFS (network file system) that can be mounted on many EC2

Answer 50

EFS

Question 51

Data Warehouse vs Data Lake. Warehouse is ________________. Lake is ___________

Answer 51

Structured, Unstructured

Question 52

Binary format that stores both the data and its schema

Answer 52

AVRO

Question 53

Columnar storage format optimized for analytics.

Answer 53

Parquet

Question 54

Find K “nearest” (most similar) rows and average their values

Answer 54

K Nearest Neighbor (KNN)

Question 55

Find linear or non-linear relationships between the missing feature and other features

Answer 55

Regression

Question 56

Duplicate samples from the minority class

Answer 56

Oversampling

Question 57

Instead of creating more positive samples, remove negative ones

Answer 57

Undersampling

Question 58

measures how “spread-out” the data is

Answer 58

Variance

Question 59

________________ 𝜎 is just the square root of the variance.

Answer 59

Standard Deviation

Question 60

Data points that lie more than one ___________________ from the mean can be considered unusual.

Answer 60

Standard Deviation

Question 61

Bucket observations together based on ranges of values.

Answer 61

Binning

Question 62

Create “buckets” for every category * The bucket for your category has a 1, all others have a 0

Answer 62

One Hot Encoding

Question 63

______________ for deploying to edge devices

Answer 63

SageMaker NEO

Question 64

___________values are the algorithm used to determine the contribution of each feature toward a model’s predictions

Answer 64

Shapley

Question 65

Used on the final output layer of a multi-class classification problem

Answer 65

Softmax

Question 66

Choosing an activation function: For multiple classification, use _________on the output layer

Answer 66

softmax

Question 67

Choosing an activation function: ________do well with Tanh

Answer 67

RNN’s

Question 68

Choosing an activation function: For everything else

Answer 68

Start with ReLU

Question 69

Choosing an activation function: _________for really deep networks

Answer 69

Swish

Question 70

When you have data that doesn’t neatly align into columns * Images that you want to find features within * Machine translation * Sentence classification * Sentiment analysis

Answer 70

Convlution Neural Network (CNN)

Question 71

RNN’s: what are they for?

Answer 71

Time-series data

Question 72

When you want to predict future behavior based on past behavior

Answer 72

Recurrent Neural Network (RNN)

Question 73

Sequence to sequence, Sequence to vector, Vector to sequence, Encoder -> Decoder

Answer 73

RNN topologies

Question 74

_________batch sizes tend to not get stuck in local minima

Answer 74

Small

Question 75

____________batch sizes can converge on the wrong solution at random

Answer 75

Large

Question 76

_________learning rates can overshoot the correct solution

Answer 76

Large

Question 77

____________learning rates increase training time

Answer 77

Small

Question 78

* ________________techniques are intended to prevent overfitting.

Answer 78

Regularization

Question 79

Preventing overfitting in ML in general * A regularization term is added as weights are learned

Answer 79

L1 and L2 Regularization

Question 80

* L1: sum of _______________

Answer 80

weights

Question 81

L2: sum of ______________

Answer 81

square of weights

Question 82

We need to understand true positives and true negative, as well as false positives and false negatives.

Answer 82

confusion matrix

Question 83

Percent of positives rightly predicted

Answer 83

Recall

Question 84

AKA Correct Positives

Answer 84

Precision

Question 85

Plot of true positive rate (recall) vs. false positive rate at various threshold settings.

Answer 85

ROC Curve

Question 86

The area under the ROC curve is… wait for it..

Answer 86

AUC

Question 87

Generate N new training sets by random sampling with replacement

Answer 87

Bagging

Question 88

Training is sequential; each classifier takes into account the previous one’s success.

Answer 88

Boosting

Question 89

Define the hyperparameters you care about and the ranges you want to try, and the metrics you are optimizing for

Answer 89

Automatic Model Tuning

Question 90

Don’t optimize too many hyperparameters at once * Limit your ranges to as small a range as possible * Use logarithmic scales when appropriate * Don’t run too many training jobs concurrently * This limits how well the process can learn as it goes * Make sure training jobs running on multiple instances report the correct objective metric in the end

Answer 90

Automatic Model Tuning: Best Practices

Question 91

Stop training in a tuning job early if it is not improving the objective significantly

Answer 91

Early Stopping

Question 92

Uses one or more previous tuning jobs as a starting point

Answer 92

Warm Start

Question 93

Automates: * Algorithm selection * Data preprocessing * Model tuning * All infrastructure * It does all the trial & error for you

Answer 93

SageMaker Autopilot

Question 94

Visual IDE for machine learning

Answer 94

SageMaker Studio

Question 95

Create and share Jupyter notebooks with SageMaker Studio * Switch between hardware configurations (no infrastructure to manage)

Answer 95

SageMaker Notebooks

Question 96

Organize, capture, compare, and search your ML jobs

Answer 96

SageMaker Experiments

Question 97

Saves internal model state at periodical intervals * Gradients / tensors over time as a model is trained * Define rules for detecting unwanted conditions while training * A debug job is run for each rule you configure * Logs & fires a CloudWatch event when the rule is hit

Answer 97

SageMaker Debugger

Question 98

Catalog your models, manage model versions * Associate metadata with models * Manage approval status of a model * Deploy models to production * Automate deployment with CI/CD

Answer 98

SageMaker Model Registry

Question 99

___________________ is a visualization toolkit for Tensorflow or PyTorch * Visualize loss and accuracy * Visualize model graph * View histograms of weight, biases over time * Project embeddings to lower dimensions * Profiling

Answer 99

Tensorboard

Question 100

Compile & optimize training jobs on GPU instances * Can accelerate training up to 50% * Converts models into hardware-optimized instructions * Tested with Hugging Face transformers library, or bring your own model * Incompatible with SageMaker distributed training libraries

Answer 100

SageMaker Training Compiler

Question 101

Retain and re-use provisioned infrastructure * Useful if repeatedly training a model to speed things up * Use by setting KeepAlivePeriodInSeconds in your training job’s resource config

Answer 101

Warm Pools

Question 102

Creates snapshots during your training * You can re-start from these points if necessary * Or use them for troubleshooting, to analyze the model at different points * Automatic synchronization with S3 (from /opt/ml/checkpoint)

Answer 102

Checkpointing

Question 103

Run automatically when using ml.g or ml.p instance types * Replaces any faulty instances * Runs GPU health checks * Ensures NVidia Collective Communication Library is working

Answer 103

Cluster Health Checks and Automatic Restarts

Question 104

You can of course run multiple training jobs in parallel * “job parallelism” * Individual training can also be parallelized * Distributed data parallelism * Distributed model parallelism

Answer 104

Distributed Training

Question 105

Network device attached to your SageMaker instances * Makes better use of your bandwidth * Promises performance of an onpremises High Performance Computing (HPC) cluster in the cloud

Answer 105

Elastic Fabric Adapter (EFA)

Question 106

______________ produces a weighted average of all token embeddings. The magic is in computing the attention weights.

Answer 106

Self-attention

Question 107

A mask can be applied to prevent tokens from “peeking” into future tokens (words)

Answer 107

Masked Self-Attention

Question 108

Chat! * Question answering * Text classification * i.e., sentiment analysis * Named entity recognition * Summarization * Translation * Code generation * Text generation * i.e., automated customer service

Answer 108

Applications of Transformers

Question 109

Tokenization, token encoding * Token embedding * Captures semantic relationships between tokens, token similarities * Positional encoding * Captures the position of the token in the input relative to other nearby tokens * Uses an interleaved sinusoidal function so it works on any length

Answer 109

LLM Input processing

Question 110

The stack of decoders outputs a vector at the end * Multiply this with the token embeddings * This gives you probabilities (logits) of each token being the right next token (word) in the sequence

Answer 110

LLM Output processing