AWS Services Flashcards
What is AWS Augmented AI ?
When a model is consistently providing low confidence results you can have human input to verify those predictions and these can be your own employees, aws staff or mechanical turk members. Some vendors are pre-screened for confidentiality requirements.
The results can be fed back into the training model.
What is AWS comprehend ?
For NLP. Uses machine learning to find insights and relationships in text looking at
1. Language of the text
2. Extracts key phrases, places, people, brands or events
3. Understands how positive or negative the text is
4. Analyzes text using tokenisation and parts of speech
5. Automatically organises a collection of text files by topic
What is Named Entity Recognition in Comprehend ?
NER extracts predefined, general purpose entities like people, places, organisations, dates and other standard categories from text. You can add custom entities as well such as policy numbers.
What is Amazon Forecast ?
Forecasting service - reduces forecasting time
You upload to S3 bucket for example your sales data which forecast will use to generate a forecasting model.
What is Amazon Kendra ?
A document search service. It extract answers from the data sources that you specify and offers natural language search capabilities.
It learns from the interactions/feedback to promote preferred results (incremental learning)
You also have the ability to manually fine tune the results.
What is Amazon Lex ?
Used to build chatbots quickly for your applications using voice and text.
Supports multiple languages
Integrates with AWS Lambda, Comprehend and Kendra.
The bot automatically understands the user intent to invoke the correct Lambda function to fulfill the intent.
The bot will ask for “Slots” or input parameters if necessary.
What is Amazon Mechanical Turk ?
A human workforce service spread internationally. You can set the reward for simple tasks. You could use this workforce to label data, act as a feedback solution for AI results.
Use Cases
Image classification
Data collection
Business processing
What is Amazon Personalise ?
Fully managed ML service to build app with real time personalised recommendations. It is the same technology used by Amazon.com.
Integrates into existing websites, applications, sms, email marketing systems. Takes days not months to build and train the model. It is the same technology used by Amazon.com.
Recipes
Out of the box use cases but still need your data. An example POPULAR_ITEMS, PERSONALIZED_RANKING, RELATED_ITEMS, USER_SEGMENTATION, USER_PERSONALISATION
These are used for recommendations.
What is amazon polly ?
Truns text into speech using deep learning. Allowing you to create applications that talk.
Lexicons - Rules that guide how words are turned into speech AWS should be pronounces Amazon Web Services.
SSML - Speech Synthesis Markup Language - Allows you to add text tags that influences speech such as <whisper> or <break></break></whisper>
Speech mark - Encode where a sentence/work starts or ends in the audio. This is helpful for lip syncing or highlighting of words as they are spoken.
What is Amazon recognition ?
Its mainly used to find objects, people, text, scenes in images and videos using ML. Facial analysis and facial search to do user verification, people counting.
You can create a database of familiar faces or compare against celebrities.
Use Cases
Labeling
Content Moderation
Text Detection
Face Detection and Analysis
Face Search and Verification
Celebrity Recognition
Pathing (Sports game analysis)
Thing to remember that the source data is an image.
Custom Labels
Custom labels are a way of training recognition to recognise a specific item such as your company logo. You basically train recognition by uploading images with the label you want and then recognition can use what it has learnt to pick out the data from other images on the internet or other sources.
Content Moderation
Used for automatically detecting inappropriate, unwanted or offensive content. An example would be the filtering out of harmful images in social media. You can use Amazon Augmented AI if you want to have a human review.
By using custom moderation adaptors you can use training data to recognise unique moderation items.
What is Amazon textract ?
Automatically extracts text, handwriting and data from any scanned documents using AI and ML. It can also extract data from forms and tables and document formats such as PDFs
Use Cases
Financial Services
Healthcare
Public Sector
What is Amazon Transcribe ?
Automatically extracts text, handwriting and data from any scanned documents using AI and ML. It can also extract data from forms and tables and document formats such as PDFs
Use Cases
Financial Services
Healthcare
Public Sector
