Curriculum
- 25 Sections
- 249 Lessons
- 10 Weeks
Expand all sectionsCollapse all sections
- 1. Introduction to Housekeeping8
- 1.1Introduction
- 1.2increase the speed of learning
- 1.3Preparation – AWS Machine Learning Specialty Exam
- 1.4Lab – AWS Account Setup, Free Tier Offers, Billing, Support
- 1.5Lab – Billing Alerts, Delegate Access
- 1.6Lab – Configure IAM Users, Setup Command Line Interface (CLI)
- 1.7Benefits of Cloud Computing
- 1.8AWS Global Infrastructure Overview
- 2. Sege Maker Housekeeping7
- 3. Machine Learning Concepts9
- 3.1Introduction to Machine Learning, Concepts, Terminologies
- 3.2Data Types – How to handle mixed data types
- 3.3Lab – Python Notebook Environment
- 3.4Lab – Working with Missing Data
- 3.5Lab – Data Visualization – Linear, Log, Quadratic and More
- 3.6AWS Sample Question #2
- 3.7Answer to Question #2
- 3.8AWS Sample Question #9
- 3.9Answer to Sample Question #9
- 4. Model Performance Evaluvation16
- 4.1Introduction
- 4.2Model Performance
- 4.3Downloadable Resources
- 4.4Lab – Binary Classifier Performance
- 4.5Lab – Binary Classifier – Confusion Matrix
- 4.6Lab – Binary Classifier – SKLearn Confusion Matrix
- 4.7Binary Classifier – Metrics Calculation
- 4.8Binary Classifier – Metrics Definition
- 4.9Question – Why not Model 1?
- 4.10Binary Classifier – Area Under Curve Metrics
- 4.11Lab – Multiclass Classifier
- 4.12Model Performance
- 4.13Model Performance Evaluation
- 4.14What metric is appropriate – Q&A Discussion
- 4.15AWS Sample Question #5 Answer to Question #5
- 4.16Answer to Question #5
- 5. SegeMaker Service Overview14
- 5.1Downloadable Resources
- 5.2How is AWS SageMaker different from other ML frameworks?
- 5.3Introduction to SageMaker
- 5.4Instance Type and Pricing
- 5.5Save Money on SageMaker Usage
- 5.6DataFormat
- 5.7SageMaker Built-in Algorithms
- 5.8Popular Frameworks and Bring Your Own Algorithm
- 5.9Infrastructure, Pricing, Support – Review
- 5.10AWS Sample Question #1
- 5.11Answer for Sample Question #1
- 5.12AWS Sample Question #10
- 5.13Answer for Sample Question #10
- 5.14What does a data scientist in gaming do? By Carly Taylor
- 6. SegeMaker Service and SDK Changes6
- 7. XGBoost - Gradient Boosted Trees30
- 7.1Downloadable Resources
- 7.2Introduction to XGBoost
- 7.3Lab – Data Preparation Simple Regression
- 7.4Lab – Training Simple Regression
- 7.5Lab – Data Preparation Non-linear Data set
- 7.6Lab – Training Non-linear Data set
- 7.7Exercise – Improving quality of predictions
- 7.8Lab – Data Preparation Bike Rental Regression
- 7.9Lab – Train Bike Rental Regression Model
- 7.10Lab – Train using Log of Count
- 7.11ResourceLimitExceeded Error – How to Increase Resource Limit
- 7.12Lab – How to train using SageMaker’s built-in XGBoost Algorithm
- 7.13Q&A: How does SageMaker built-in know the target variable?
- 7.14Lab – How to run predictions against an existing SageMaker Endpoint
- 7.15Q&A – XGBoost on SageMaker predicted values are not delimited consistently
- 7.16SageMaker Endpoint Features
- 7.17SageMaker Spot Instances – Save up to 90% for training jobs
- 7.18Lab – Multi-class Classification
- 7.19Lab – Binary Classification
- 7.20Exercise – Improve Data Quality in Diabetes dataset
- 7.21Question on Diabetes Data Quality Improvement
- 7.22Question on Diabetes model – is group mean on target the right approach?
- 7.23Data Leakage
- 7.24Exercise – Mushroom Classification
- 7.25Quiz – XGBoost
- 7.268 questions
- 7.27Underfitting, Overfitting
- 7.283 questions
- 7.29AWS Sample Question #8
- 7.30Answer to AWS Sample Question #8
- 8. Invoke Model Endpoint from External Clients9
- 8.1Install SageMaker SDK, GIT Client, Source Code, Security Permissions
- 8.2IAM users for the lab Integration Overview
- 8.3Lab – Client to Endpoint using SageMaker SDK
- 8.4Lab – Client to Endpoint using Boto3 SDK
- 8.5Microservice – Lambda to Endpoint – Payload
- 8.6Lambda UI Changes
- 8.7Lab – Microservice – Lambda to Endpoint
- 8.8API Gateway – UI Changes
- 8.9Lab – API Gateway, Lambda, Endpoint
- 9. EndPoint Changes with Zero Downtime0
- 10. Emerging Ai Trends and Social issues3
- 11. Cloud Security and Access Management12
- 11.1Introduction
- 11.2Shared Responsibility Model, Compliance, Delegation, Federation
- 11.3Credentials, MFA, Identity-based, Resources-based Policy
- 11.4Inline and Managed Policy, Amazon Resource Naming (ARN) Convention
- 11.5Principal, Effect, Action, Resource, Not Clause
- 11.6Conditional Access, Implicit Deny, Explicit Allow and Deny, Permission Boundary
- 11.7IAM Roles, Cross-account access options
- 11.8Federation, SSO, SAML, Active Directory, AWS Organizations, Cognito
- 11.9Lab – Identity-based policy, Implicit Deny, Explicit Allow
- 11.10Lab – Policy Generator, Managed Policy, Versions, Groups
- 11.11Lab – Resource-based policy, Policy Generator, Principals
- 11.12Cloud Security
- 12. Principal Component Analysis (PCA)13
- 12.1Normalization and Standardization
- 12.2Downloadable Resources
- 12.3Introduction to Principal Component Analysis
- 12.4PCA Demo Overview
- 12.5Demo – PCA with Random Dataset
- 12.6Demo – PCA with Correlated Dataset
- 12.7Cleanup Resources on SageMaker
- 12.8Demo – PCA with Kaggle Bike Sharing – Overview and Normalization
- 12.9Demo – PCA Local Mode with Kaggle Bike Train
- 12.10Demo – PCA training with SageMaker
- 12.11Demo – PCA Projection with SageMaker
- 12.12Exercise : Kaggle Bike Train and PCA
- 12.13Summary
- 13. Recommender Systems - Factorization Machines7
- 14. Model Optimization and HyperParameter Tuning8
- 14.1Downloadable Resources
- 14.2Introduction to Hyperparameter Tuning
- 14.3Lab: Tuning Movie Rating Factorization Machine Recommender System
- 14.4Lab: Step 2 Tuning Movie Rating Recommender System
- 14.5HyperParameter, Bias-Variance, Regularization (L1, L2) [Repeat from XGBoost]
- 14.6Nuts and Bolts of Optimization
- 14.6Model Optimization
- 14.6Model Optimization – related question
- 15. Time Series Forecasting - DeepAR12
- 15.1Downloadable Resources
- 15.2Introduction to DeepAR Time Series Forecasting
- 15.3DeepAR Training and Inference Formats
- 15.4Working with Time Series Data, Handling Missing Values
- 15.5Demo – Bike Rental as Time Series Forecasting Problem
- 15.6Demo – Bike Rental Model Training
- 15.7Demo – Bike Rental Prediction
- 15.8Demo – DeepAR Categories
- 15.9Demo – DeepAR Dynamic Features Data Preparation
- 15.10Demo – DeepAR Dynamic Features Training and Prediction
- 15.11Summary
- 15.12Question: How to train a model for different products using DeepAR?
- 16. Anomaly Detection - Random Cut Forest4
- 17. Artificial Intelligence (A) Services13
- 17.1Downloadable Resources
- 17.2Lab Instructions
- 17.31. Introduction
- 17.42.1 Amazon Transcribe and Lab
- 17.52.2 Amazon Transcribe and Lab
- 17.63. Amazon Translate Translate – Practical Scenario
- 17.74.1 Amazon Comprehend Pricing Comprehend
- 17.84.2 Amazon Comprehend
- 17.94.3 Amazon Comprehend training
- 17.105. Amazon Polly
- 17.117. Amazon Rekognition
- 17.126. Amazon Lex
- 17.138. Amazon Textract & Summary AI Services Quiz
- 18. S3 Data Lake Architecture - Data Consolidation22
- 18.1Downloadable Resources
- 18.2Lab Instructions
- 18.3Introduction to Data Lake
- 18.4Kinesis – Streaming and Batch Processing
- 18.5Data Formats and Tools for Data Format Conversion
- 18.6In-Place Analytics and Portfolio of Tools
- 18.7Monitoring and Optimization
- 18.8Security and Protection
- 18.9Quiz – Data Lake
- 18.10Lab Instructions – Glue Data Catalog
- 18.11Lab – Glue Data Catalog
- 18.12Lab Instructions – Athena In-place Querying
- 18.13Lab – Query with Athena
- 18.14Lab – Glue ETL – Convert format to Parquet
- 18.15Lab – Query Amazon Customer Reviews with Athena
- 18.16Lab – Sentiment of the Customer Review
- 18.17Lab – Query Sentiment of Customer Reviews using Athena
- 18.18Lambda UI Changes
- 18.19Lab – Serverless Customer Review Solution Part 1
- 18.20Lab – Serverless Customer Review Solution Part 2
- 18.21AWS Sample Question #3
- 18.22Answer to Sample Question #3
- 19. Deep Learning and Neural Networks26
- 19.1ReadMe and Downloadable Resources
- 19.2Regression – Gradient Descent Batch, Mini-Batch, Stochastic, Loss, RMSProp, Adam
- 19.3Classification – Gradient Descent, Loss Function
- 19.4Neural Networks and Deep Learning
- 19.5Real World Face Restoration
- 19.6Lab – Regression with SKLearn Neural Network
- 19.7Lab – Regression with Keras and TensorFlow
- 19.8Customer Churn Data
- 19.9Lab – Binary Classification – Part 1- Customer Churn Prediction
- 19.10Lab – Binary Classification – Part 2 – Customer Churn Prediction
- 19.11Lab – Multiclass Classification – Iris
- 19.12Transfer Learning
- 19.13Optimizing for GPUs
- 19.14Multi-Class Multi-Label Classification
- 19.15Nuts and Bolts of Optimization [Repeat]
- 19.16Quiz – Neural Network and Model Tuning
- 19.17Answer to Sample Question #4
- 19.18AWS Sample Question #4
- 19.19AWS Sample Question #6
- 19.20Answer to Sample Question #6
- 19.21AWS Sample Question #7
- 19.22Answer to Sample Question #7
- 19.23MIT – Introduction to Deep Learning
- 19.24Convolutional Neural Network (CNN)
- 19.25Recurrent Neural Networks (RNN), LSTM
- 19.26Generative Adversarial Networks (GANs)
- 20. Bring Your Own Algorithm8
- 20.1How to use TensorFlow, Pytorch, SKLearn in SageMaker
- 20.2Downloadable Resources
- 20.3Introduction and How built-in algorithms work
- 20.4Custom Image and Popular Framework
- 20.5Folder Structure and Environment Variables
- 20.6Lab – SKLearn Estimator Bring Your Own Part 1
- 20.7Lab – SKLearn Estimator Bring Your Own Part 2
- 20.8Lab – TensorFlow Estimator Bring Your Own
- 21. Storage for Servers5
- 22. AWS - Support Plans and Feedback.2
- 23. Databases on AWS8
- 23.1Downloadable Resources
- 23.2AWS Databases – Introduction, Benefits, and Types
- 23.3Relational Database Service (RDS) – Features and Benefits
- 23.4Aurora and Aurora Serverless Relational Database
- 23.5DynamoDB – Primary Key, Partitions, and Features
- 23.6Cassandra and DocumentDB
- 23.7Amazon ElastiCache – Usage Example, Features
- 23.8Amazon Redshift
- 24. On-Premises usage and other technologies3
- 25. Practice Exam - AWS Certified Machine Learning Specialty4
