ExamPark
Course menuAIF-C01

Exams / AIF-C01 / Study guide / AWS managed AI services

AWS managed AI services: AIF-C01 study notes and practice questions

What to know about AWS managed AI services for the AIF-C01 exam, from the Fundamentals of AI and ML domain (20% of the exam guide), then 2 free questions to test it. Part of the AIF-C01 study guide.

Last verified against current AWS documentation ·How we write our questions

What to know for the exam

These questions name a task and ask for the AWS service that does it. The pre-trained AI services need no training data or ML skills, so they are the lowest-effort answer when they fit.

The services

Task Service
Find objects, faces, text or unsafe content in images and video Amazon Rekognition
Extract text, forms and tables from scanned documents Amazon Textract
Understand text: sentiment, entities, key phrases, language, PII Amazon Comprehend (Amazon Comprehend Medical for health text)
Speech to text Amazon Transcribe
Text to lifelike speech Amazon Polly
Translate text between languages Amazon Translate
Build voice and text chatbots Amazon Lex
Intelligent search over company documents Amazon Kendra
Personalized recommendations Amazon Personalize
Detect online fraud Amazon Fraud Detector
Human review of ML predictions Amazon Augmented AI (A2I)
Analyze customer contact center calls Contact Lens for Amazon Connect

Building your own models

Amazon SageMaker AI is the platform to build, train and deploy custom ML models. Parts that come up on the exam:

  • SageMaker Canvas: no-code ML for business analysts.
  • SageMaker JumpStart: pre-trained models, including open foundation models, you can deploy and fine-tune.
  • SageMaker Clarify: detect bias and explain predictions.
  • SageMaker Model Monitor: watch deployed models for drift.
  • SageMaker Ground Truth: label training data, with human workers.
  • SageMaker Data Wrangler: prepare data.

Common traps

  • SageMaker AI when a ready-made service fits. Sentiment analysis is Comprehend, not a custom model.
  • Textract vs Comprehend. Textract reads text out of documents; Comprehend understands text.
  • Transcribe vs Polly. Transcribe listens; Polly speaks.
  • Kendra vs Bedrock Knowledge Bases. Kendra is enterprise search; a knowledge base feeds retrieved text to a foundation model to generate an answer.

Try the 2 practice questions

Practice AWS managed AI servicesTake the free practice exam

Question 1

Medium ·Link to this question (question 1)

A media and publishing group is choosing AWS AI services for five jobs.

Match each job to the managed service that does it. Each option is used at most once, and one option is not needed.

Match each item to one answer. An answer can be used more than once, and some may not be used.

Items

  1. Turn the audio from a recorded podcast series into searchable written text
  2. Read article summaries aloud in a natural-sounding voice for a listening app
  3. Find the sentiment and main topics in thousands of reader comments
  4. Extract text, tables and form fields from scanned contracts and invoices
  5. Publish the same article in Spanish and German on the same day

Answers

  1. A. Amazon Comprehend, which finds sentiment, key phrases and entities in text.
  2. B. Amazon Lex, which builds conversational bots for voice and text.
  3. C. Amazon Polly, which turns written text into lifelike speech.
  4. D. Amazon Textract, which extracts text, tables and form fields from scanned documents.
  5. E. Amazon Transcribe, which converts speech in audio and video files into text.
  6. F. Amazon Translate, which translates text between languages.
Show the answer and explanation

Correct matches

  • Turn the audio from a recorded podcast series into searchable written text: E. Amazon Transcribe, which converts speech in audio and video files into text.
  • Read article summaries aloud in a natural-sounding voice for a listening app: C. Amazon Polly, which turns written text into lifelike speech.
  • Find the sentiment and main topics in thousands of reader comments: A. Amazon Comprehend, which finds sentiment, key phrases and entities in text.
  • Extract text, tables and form fields from scanned contracts and invoices: D. Amazon Textract, which extracts text, tables and form fields from scanned documents.
  • Publish the same article in Spanish and German on the same day: F. Amazon Translate, which translates text between languages.

About each answer

  • A. Comprehend uses natural language processing to read text and return sentiment, key phrases and entities. It matches the reader comment need without a custom model.
  • B. Lex runs conversations with users in voice or text. None of the five jobs is a dialogue, so this is the option that is not used.
  • C. Polly converts written text into natural-sounding speech. It matches the need to read article summaries aloud in a listening app.
  • D. Textract reads scanned documents and returns their text, tables and form fields. It matches the need to pull data out of contracts and invoices.
  • E. Transcribe turns recorded speech into text, which suits the podcast archive. It matches the need to make the recordings searchable.
  • F. Translate converts text from one language into another. It matches the need to publish the same article in Spanish and German.

Explanation

Each AWS AI service handles one kind of input or output. Transcribe turns speech into text, Polly turns text into speech, Comprehend analyzes the meaning of text, Textract extracts data from documents, and Translate converts between languages. Lex builds dialogues, which none of the five jobs needs. This question tests matching each job to the service built for it.

Study

Question 2

Easier ·Link to this question (question 2)

A healthcare analytics firm's data science team wants to build, train and deploy its own machine learning models on its own data.

It needs managed notebooks, managed training and managed hosting for each step of the workflow, without running servers.

Which AWS service is designed for this?

Choose one answer.

  1. A. Amazon Bedrock, which gives API access to ready-made foundation models.
  2. B. Amazon Comprehend, which analyzes text with prebuilt language models.
  3. C. Amazon SageMaker AI, which builds, trains and deploys custom ML models.
  4. D. Amazon Rekognition, which analyzes images with prebuilt vision models.
Show the answer and explanation

Answer: C

  • A. Wrong: Bedrock is aimed at using foundation models through an API. It is not built to run the full workflow for the team's own custom models.
  • B. Wrong: Comprehend uses prebuilt models to analyze text. It does not let the team build and deploy its own models on its own data.
  • C. Right: Correct. SageMaker AI is a fully managed ML service where data scientists build, train and deploy models into a hosted environment. It covers every step of the workflow. Learn more: What is Amazon SageMaker AI? (opens in a new tab)
  • D. Wrong: Rekognition analyzes images with prebuilt models. It does not give the team a workflow for training models on its own data.

Explanation

Amazon SageMaker AI is the service for building, training and deploying custom machine learning models, with managed infrastructure for each step. The other options offer prebuilt AI, either as an API to foundation models or as ready-made analysis for text or images. This question tests knowing which service is for custom ML.

Study

Want more on AWS managed AI services? The 5 full-length practice sets have 14 more AWS managed AI services questions, 325 in all, none of them in the free exam. One payment of $14.99, no subscription. See the practice sets

Study AWS managed AI services

More Fundamentals of AI and ML topics