Graduate Program KB

Section 18 - Machine Learning

Rekognition

  • Find objects, people, text, scenes in images and videos
  • Use cases:
    • Labeling, content moderation, text detection, face detection and analysis, face search and verification, celebrity recognition, pathing
  • Content moderation
    • Detect inappropriate, unwanted or offensive content
    • Used commonly in media, advertising, e-commerce scenarios to create a safe user experience and comply with regulations
    • Set a minimum confidence threshold for items that will be flagged
    • Flag sensitive content for manual review in Amazon Augmented AI

Transcribe

  • Converts speech to text via a deep learning process called automatic speech recognition
  • Automatically removes personally identifiable information using redaction
  • Supports automatic language identification for multilingual audio
  • Useful for transcribing customer service calls, automate closed captioning and subtitling, generating metadata for media assets to create a fully searchable archive

Polly

  • Turn text into speech via deep learning
  • Can create applications that talk
  • Lexicon & SSML
    • Customize pronunciation with pronunciation lexicons
    • Upload lexicons and use them in SynthesizeSpeech operation
    • Generate speech from plain text or from documents marked up with Speech Synthesis Markup Language which enables more customizations
      • Emphasize specific words or phrases
      • Use phonetic pronunciation
      • Include breathing sounds and whispering
      • Using the Newscaster speaking style

Translate

  • Language translation
  • Allows you to localize content such as websites and apps so international users can easily translate large volumes of text efficiently

Lex & Connect

  • Lex:
    • Uses ASR to convert speech to text
    • Has natural language understanding to recognize intent of text and callers
    • Helps build chatbots and call centre bots
  • Connect:
    • Receive calls, create contact flows, cloud-based virtual contact centre
    • Can integrate with other CRM systems or AWS
    • No upfront payments, 80% cheaper than traditional contact centre solutions

Comprehend

  • A fully managed and serverless service for natural language processing
  • Uses ML to find insights and relationships in text
    • Language
    • Phrases, places, people, brands, events
    • Understands the positivity or negativity of the text
    • Analyze text using tokenization
    • Automatically organizes a collection of text files by topic
  • Useful for analyzing customer interactions to find out about experiences, creating and grouping articles by topics that Comprehend will uncover
  • Comprehend Medical
    • Detects and returns useful information in unstructured clinical text (physician's notes, discharge summaries, test results, case notes)
    • Uses NLP to detect protected health information
    • Stores documents in S3, analyzes real-time data with Firehouse or use Transcribe to transcribe patient narratives into text analyzable by Comprehend Medical

SageMaker

  • A fully managed service for developers or data scientists to build ML models

Forecast

  • A fully managed service using ML to deliver accurate forecasts
  • Reduce forecasting time from months to hours
  • Useful for product demand planning, financial planning, resource planning, etc.

Kendra

  • A fully managed document search service using ML
  • Extracts answers from within a document using natural language search capabilities
  • Leverages incremental learning, learning from user interactions to promote preferred results
  • Has the ability to manually fine-tune search results

Personalize

  • A fully managed service to build apps with real-time personalized recommendations
  • Can integrate into existing websites, apps, SMS, email marking systems, etc.
  • Implement in days as opposed to months, not need to build, train and deploy
  • Useful for retail stores, media and entertainment

Textract

  • Extracts text, handwriting and data from any scanned documents
  • Can read and process any type of document
  • Useful for financial services, healthcare, public sector, etc.