AWS Managed AI Services Flashcards
What AWS services are used for text and document processing?
Amazon Comprehend, Amazon Translate, Amazon Textract.
What AWS service is used for image and video analysis?
Amazon Rekognition.
What AWS service is used for search?
Amazon Kendra.
What AWS service is used for building chatbots?
Amazon Lex.
What AWS services are used for speech processing?
Amazon Polly (text-to-speech) and Amazon Transcribe (speech-to-text).
What AWS service is used for personalized recommendations?
Amazon Personalize.
What AWS service provides a full machine learning platform?
Amazon SageMaker.
What is Amazon Comprehend used for?
Natural Language Processing (NLP).
What insights can Amazon Comprehend extract from text?
Key phrases, places, people, brands, events, and sentiment analysis.
What is custom classification in Amazon Comprehend?
A feature that allows users to define and train custom document categories.
What is Named Entity Recognition (NER) in Amazon Comprehend?
A feature that extracts predefined entities like people, places, organizations, and dates.
What is custom entity recognition in Amazon Comprehend?
A feature that allows training the model to identify specific terms, such as policy numbers. or phrases like escalation etc.
What does Amazon Transcribe do?
It converts speech into text using automatic speech recognition (ASR).
What is ASR in Amazon Transcribe?
Automatic Speech Recognition – a deep learning technique to convert speech to text.
Can Amazon Transcribe remove sensitive data?
Yes, it supports automatic redaction of Personally Identifiable Information (PII).
How can you improve recognition of domain-specific words in Amazon Transcribe?
By using custom vocabularies.
How are custom vocabularies different from custom language models?
Custom vocabularies focus on specific words; custom language models focus on context.
What feature does Amazon Transcribe offer to detect harmful speech?
Toxicity detection.
What data types does toxicity detection use?
Speech cues (tone and pitch) and text-based cues (like profanity or hate speech).
What are some categories of toxicity in Amazon Transcribe?
Sexual harassment, hate speech, threats, abuse, profanity, insult, graphic.
How can you maximize transcription accuracy in Amazon Transcribe?
Use both custom vocabularies and custom language models.
What does Amazon Polly do?
It converts text into lifelike speech using deep learning.
What do lexicons in Amazon Polly do?
They allow custom pronunciation of specific words or acronyms (e.g., AWS as Amazon Web Services).
What is SSML in Amazon Polly?
Speech Synthesis Markup Language – it controls pronunciation, pauses, emphasis, and tone.
What can SSML be used for in Polly?
Adding breaks, whispering, emphasizing words, and controlling pronunciation.
What are the different voice engines in Amazon Polly?
Neural, standard, long-form, and generative.
Which Amazon Polly voice engine provides the most human-like speech?
Generative.
What are speech marks in Amazon Polly?
They indicate the timing of words in speech for applications like lip-syncing or word highlighting.
