Machine Learning Flashcards
What is Rekognition?
A service for finding people, text and scenes, objects in images and videos, using ML.
What are rekognition features?
facial recognition for user verification, or people counting.
create a database of familiar faces.
Face analysis (gender, age, emotions).
Content moderation. (Inappropiate, unwanted, offensive).
Text detection.
Pathing.
What is aws transcribe?
Convert speech to text
It has redact feature to remove personal info from audio to text
It has auto language identification
What is Polly?
Opposite of transcribe. Turn text into speech with deep learning.
You can customize pronunciation of certain words.
You can use the SSML language, which is for customizing how the voice will speak.
What is AWS Translate?
Translate large volumes of text efficiently.
What is Amazon LEX?
Same technology that power Alexa.
It helps you build chatbots, or call center bots.
What is amazon connect?
Its a Virtual Contact Center: Receive calls, etc.
What is amazon Comprehend?
Its for natural language processing: NLP
Serverless service, that uses ML to find stuff in text.
Stuff like language, key information, intent.
TLDR: Tries to understand the meaning of the data in processes.
What is a comprehend use case?
Analyze customer interactions, like emails, to find out what leads to positive or negative experience. For improvement.
What is comprehend medical?
NLP that analyzes unstructured medical texts and and returns useful structured info.
What is SageMaker?
Fully managed service for developer, to build machine learning models.
With SageMaker, you can train a ML model, giving it historical data, so it learns to predict outcomes when you give new data to it.
What is AWS Forecast?
Fully managed service that uses ML to deliver highly accurate forecasts of anything.
Example: Predict the future sales of a product.
What is Kendra?
ML Service to search documents.
You upload any document to kendra, it builds an index of all the data in the document, and you can ask questions to it to get answers from that data.
Similar to what google search does with the internet.
What is amazon personalize?
ML Service to make personalized recommendations based on your preferences in apps.
For example when you search something in a website like amazon.com, then it recommends you similar stuff to what you searched.
What is Textract?
Extracts text data from text, tables, or scanned documents.
For example upload a drivers license, textract will analyze it and give you the result data in a text document.
What are all the ML services for the exam?
Rekognition
Transcribe
Polly
Translate
Lex
Connect
Comprehend
SageMaker
Forecast
Kendra
Personalize
Which fully managed service can deliver highly accurate forecasts?
Forecast
You would like to find objects, people, text, or scenes in images and videos. What AWS service should you use?
Rekognition
A start-up would like to rapidly create customized user experiences. Which AWS service can help?
Personalize
Amazon Personalize is a machine learning service that makes it easy for developers to create individualized recommendations for customers using their applications.
A research team would like to group articles by topics using Natural Language Processing (NLP). Which service should they use?
Comprehend
Amazon Comprehend is a natural language processing (NLP) service that uses machine learning to find meaning and insights in text.
A company would like to convert its documents into different languages, with natural and accurate wording. What should they use?
Translate
A developer would like to build, train, and deploy a machine learning model quickly. Which service can he use?
SageMaker
Which AWS service makes it easy to convert speech-to-text?
Transcribe
Which of the following services is a document search service powered by machine learning?
Kendra
Which ML Service is similar to google but for documents?
Kendra
A company is managing an image and video sharing platform which is used by customers around the globe. The platform is running on AWS using an S3 bucket to host both images and videos and using CloudFront as the CDN to deliver content to customers all over the world with low latency. In the last couple of months, a lot of customers have complained that they have started to see inappropriate content on the platform which started to increase in the last week. It will be very expensive and time-consuming to manually approve those images and videos by employees before its published on the platform. There is a requirement to find a solution that can automatically detect inappropriate and offensive images and videos and give you the ability to set a minimum confidence threshold for items that will be flagged and allows for manual review. Which AWS service can fit the requirement?
Rekognition
An online medical company that allows you to book an appointment with doctors using through a phone call is using AWS to host their infrastructure. They are using Amazon Connect and Amazon Lex to receive calls and create a workflow, book an appointment, and pay. According to the company’s policy, all calls must be recorded for review. But, there is a requirement to remove any Personally Identifiable Information (PII) from the call before it’s saved. What do you recommend to use which helps in removing PII from calls?
Transcribe
Amazon Polly allows you to turn text into speech. It has two important features. First is ……………….. which allows you to customize the pronunciation of words (e.g., “Amazon EC2” will be “Amazon Elastic Compute Cloud”). The second is ……………….. which allows you to emphasize words, including breathing sounds, whispering, and more.
1) Speech Synthesis Markup Language (SSML), Pronunciation Lexicons
2) Pronunciation Lexicons, Security Assertion Markup Language (SAML)
3) Pronunciation Lexicons, Speech Synthesis Markup Language (SSML)
4) Security Assertion Markup Language (SAML), Pronunciation Lexicons
3) Pronunciation Lexicons, Speech Synthesis Markup Language (SSML)
A medical company is in the process of implementing a solution to detect, extract, and analyze information from unstructured medical text like doctors’ notes, clinical trial reports, and radiology reports. Those documents are uploaded and stored on S3 buckets. According to the company’s regulations, the solution must be designed and implemented to keep patients’ privacy by identifying Protected Health Information (PHI) so the solution will be eligible with HIPAA. Which AWS service should you use?
Amazon Comprehend Medical
You should use Amazon Transcribe to turn text into lifelike speech using deep learning.
True or False?
False.
Amazon Transcribe is an AWS service that makes it easy for customers to convert speech-to-text. Amazon Polly is a service that turns text into lifelike speech.
A company would like to implement a chatbot that will convert speech-to-text and recognize the customers’ intentions. What service should it use?
Lex.
Amazon Lex is a service for building conversational interfaces into any application using voice and text. Lex provides the advanced deep learning functionalities of automatic speech recognition (ASR) for converting speech to text, and natural language understanding (NLU) to recognize the intent of the text, to enable you to build applications with highly engaging user experiences and lifelike conversational interactions.
Transcribe doesn’t have intention recognition.
