AWS Machine Learning Landscape

Levels of Machine Learning on AWS

1. AI Services

• Designed for application developers

• API-driven opinionated services

• No deep ML knowledge required

• Specific task-focused solutions

2. Training and Deployment Services

• Includes Amazon SageMaker and Ground Truth

• Designed for ML developers and data scientists

• Enables quick provisioning of managed compute

• Supports model training and hosting

3. Frameworks and Infrastructure

• Aimed at ML researchers and academics

• Includes Deep Learning AMIs and AWS Greengrass

• Supports custom ML frameworks

• Enables on-premises implementation

AI Developer Services

Common characteristics:

• Simple API integration

• No ML experience required

• Highly scalable and available

• Pay-per-use pricing model

• AWS SDK integration

Service Descriptions

1. Amazon Comprehend

   • Purpose: Text analysis and insight extraction

   • Primary use case: Sentiment analysis

   • Example: Monitoring negative product reviews on social media

2. Amazon Forecast

   • Purpose: Time-series data analysis

   • Use case: Predictive analytics

   • Example: Forecasting seasonal product demand

3. Amazon Lex

   • Purpose: Conversational interface creation

   • Primary use case: Chatbot development

   • Example: Website customer service automation

4. Amazon Personalize

   • Purpose: Recommendation engine

   • Processes demographic and behavioral data

   • Example: Product recommendations during checkout

5. Amazon Polly

   • Purpose: Text-to-speech conversion

   • Use case: Voice response generation

   • Example: Dynamic call center voice responses

6. Amazon Rekognition

   • Purpose: Image analysis

   • Capabilities: Object, people, and activity recognition

   • Example: Facial recognition for employee authentication

7. Amazon Textract

   • Purpose: Document data extraction

   • Supports: Images and PDFs

   • Example: Digitizing physical paper forms

8. Amazon Transcribe

   • Purpose: Speech-to-text conversion

   • Use case: Audio transcription

   • Example: Creating transcripts of recorded presentations

9. Amazon Translate

   • Purpose: Language translation

   • Use case: Content localization

   • Example: Automatic website translation based on user geography

Implementation Features

• All services accessible through AWS Console

• Demo environments available for testing

• Simple integration with existing applications

• Confidence scoring for predictions

• Serverless architecture

Key Exam Considerations

• Focus on AI developer services

• Understanding service purposes and use cases

• Knowledge of when to apply specific services

• Basic understanding of service capabilities

• Integration with serverless applications