Shreyas Subramanian - AWS Certified Machine Learning Study Guide

1 Cover

2 Title Page AWS Certified Machine Learning Study Guide Specialty (MLS-C01) Exam Shreyas Subramanian Stefan Natu

3 Copyright

4 Dedication

5 Acknowledgments

6 About the Authors

7 About the Technical Editor

8 Introduction The AWS Certified Machine Learning Specialty Exam Who Should Buy This Book Study Guide Features AWS Certified Machine Learning Specialty Exam Objectives

9 Assessment Test

10 Answers to Assessment Test

11 PART I: Introduction Chapter 1: AWS AI ML Stack Amazon Rekognition Amazon Textract Amazon Transcribe Amazon Translate Amazon Polly Amazon Lex Amazon Kendra Amazon Personalize Amazon Forecast Amazon Comprehend Amazon CodeGuru Amazon Augmented AI Amazon SageMaker AWS Machine Learning Devices Summary Exam Essentials Review Questions Chapter 2: Supporting Services from the AWS Stack Storage Amazon VPC AWS Lambda AWS Step Functions AWS RoboMaker Summary Exam Essentials Review Questions

12 PART II: Phases of Machine Learning Workloads Chapter 3: Business Understanding Phases of ML Workloads Business Problem Identification Summary Exam Essentials Review Questions Chapter 4: Framing a Machine Learning Problem ML Problem Framing Recommended Practices Summary Exam Essentials Review Questions Chapter 5: Data Collection Basic Data Concepts Data Repositories Data Migration to AWS Summary Exam Essentials Review Questions Chapter 6: Data Preparation Data Preparation Tools Summary Exam Essentials Review Questions Chapter 7: Feature Engineering Feature Engineering Concepts Feature Engineering Tools on AWS Summary Exam Essentials Review Questions Chapter 8: Model Training Common ML Algorithms Local Training and Testing Remote Training Distributed Training Monitoring Training Jobs Debugging Training Jobs Hyperparameter Optimization Summary Exam Essentials Review Questions Chapter 9: Model Evaluation Experiment Management Metrics and Visualization Summary Exam Essentials Review Questions Chapter 10: Model Deployment and Inference Deployment for AI Services Deployment for Amazon SageMaker Advanced Deployment Topics Summary Exam Essentials Review Questions Chapter 11: Application Integration Integration with On-Premises Systems Integration with Cloud Systems Integration with Front-End Systems Summary Exam Essentials Review Questions

13 PART III: Machine Learning Well-Architected Lens Chapter 12: Operational Excellence Pillar for ML Operational Excellence on AWS Summary Exam Essentials Review Questions Chapter 13: Security Pillar Security and AWS Secure SageMaker Environments AI Services Security Summary Exam Essentials Review Questions Chapter 14: Reliability Pillar Reliability on AWS Change Management for ML Failure Management for ML Summary Exam Essentials Review Questions Chapter 15: Performance Efficiency Pillar for ML Performance Efficiency for ML on AWS Summary Exam Essentials Review Questions Chapter 16: Cost Optimization Pillar for ML Common Design Principles Cost Optimization for ML Workloads Summary Exam Essentials Review Questions Chapter 17: Recent Updates in the AWS AI/ML Stack New Services and Features Related to AI Services New Features Related to Amazon SageMaker Summary Exam Essentials

14 Appendix Answers to the Review Questions Chapter 1: AWS AI ML Stack Chapter 2: Supporting Services from the AWS Stack Chapter 3: Business Understanding Chapter 4: Framing a Machine Learning Problem Chapter 5: Data Collection Chapter 6: Data Preparation Chapter 7: Feature Engineering Chapter 8: Model Training Chapter 9: Model Evaluation Chapter 10: Model Deployment and Inference Chapter 11: Application Integration Chapter 12: Operational Excellence Pillar for ML Chapter 13: Security Pillar Chapter 14: Reliability Pillar Chapter 15: Performance Efficiency Pillar for ML Chapter 16: Cost Optimization Pillar for ML

15 Index

16 End User License Agreement

1 Chapter 1 TABLE 1.1 Various features of SageMaker corresponding to the different phase...

2 Chapter 2TABLE 2.1 AWS Lambda limits

3 Chapter 5TABLE 5.1 Table of housing data

4 Chapter 8TABLE 8.1 Services relevant to an end-to-end machine learning workflow that ...

1 Chapter 1 FIGURE 1.1 Document analysis with human review flow FIGURE 1.2 Flow showing how to translate customer service calls followed by ... FIGURE 1.3 The AppointmentBotcan be built using Amazon Lex and backend ... FIGURE 1.4 The end-to-end flow with Amazon Personalize (text on top) and how...

2 Chapter 2FIGURE 2.1 Pattern for using FSx for Lustre with Amazon SageMaker for traini...FIGURE 2.2 Architecture showing the use of VPC endpoints to connect to vario...FIGURE 2.3 An example ML pipeline constructed using step functions that orch...

3 Chapter 3FIGURE 3.1 Diagram showing the phases of the machine learning lifecycle

4 Chapter 5FIGURE 5.1 Various data sources you can use with AWS Data Pipeline to land d...FIGURE 5.2 Various data sources you can use with AWS DMS to land data in S3...FIGURE 5.3 Conceptual diagram of Kinesis Data StreamsFIGURE 5.4 Conceptual diagram of Kinesis Data Firehose showing how data can ...FIGURE 5.5 Diagram showing streaming data flow pattern for Kinesis Data Anal...

5 Chapter 6FIGURE 6.1 Diagram showing SageMaker Ground Truth data labeling toolFIGURE 6.2 Diagram showing AWS Glue as an ETL tool

6 Chapter 7FIGURE 7.1 Diagram showing how you can deal with skewed distributions using ...FIGURE 7.2 Diagram showing how you backtest on time series data

7 Chapter 8FIGURE 8.1 Linear regression example. The error terms shown by the vertical ...FIGURE 8.2 Example showing lack of constant variance. The error terms shown ...FIGURE 8.3 Example showing violation of linearity assumptionFIGURE 8.4 SVM conceptual example showing the separatrix by the solid line a...FIGURE 8.5 You can use a kernel SVM to separate the points. Although they cl...FIGURE 8.6 Sequential learning of XGBoost to combine many weak learners into...FIGURE 8.7 Possible output of a clustering analysis that splits data into th...FIGURE 8.8 Elbow curve analysis of PCA to determine optimal number of cluste...FIGURE 8.9 Values of N1 and N2 explored with grid searchFIGURE 8.10 Values of N1 and N2 explored with random search

8 Chapter 9FIGURE 9.1 Data from a toy two-class classification problemFIGURE 9.2 Example SVM hyperplane separating the two classes of dataFIGURE 9.3 Example ROC curveFIGURE 9.4 Example showing comparison of two ROC curves by calculating the A...FIGURE 9.5 Example precision vs. recall curve

9 Chapter 10FIGURE 10.1 SageMaker real-time endpoints under the hoodFIGURE 10.2 SageMaker Batch transform under the hoodFIGURE 10.3 Re-create strategy showing how to stop endpoint A and start endp...FIGURE 10.4 Ramped strategy showing how to gradually shift from endpoint A t...

10 Chapter 11FIGURE 11.1 Typical architecture for connecting Amazon API Gateway to AWS La...

11 Chapter 12FIGURE 12.1 Diagram showing the ML workflow with different ways you can use ...FIGURE 12.2 Diagram showing a typical CI/CD workflow that can be used as par...

12 Chapter 13FIGURE 13.1 Diagram showing the AWS shared responsibility model. Understand ...FIGURE 13.2 Diagram showing tag-based controls that can be applied using att...FIGURE 13.3 Authentication flow using SAML 2.0 to access the AWS consoleFIGURE 13.4 Different IAM roles applicable to SageMakerFIGURE 13.5 Private network traffic to SageMaker Studio

1 Cover

2 Table of Contents

3 Title Page AWS Certified Machine Learning Study Guide Specialty (MLS-C01) Exam Shreyas Subramanian Stefan Natu

4 Copyright

5 Dedication

6 Acknowledgments