Recently I’ve attended a workshop on “AI on AWS” under the mentorship of Mr.Vimal Daga

It was an Expertise Level Workshop consisting of all the AI services on AWS like:

🔹Amazon Rekognition

🔹Amazon Polly

🔹Amazon Lex

🔹Amazon Kendra

🔹Amazon Forecast

🔹Amazon CodeGuru

🔹Amazon Translate

🔹Amazon Textract

🔹Amazon Comprehend

🔹Amazon Transcribe

🔹Amazon Personalize

🔹Amazon Fraud Detector

Key Takeaway:

  • ML provides the system capabilities to learn and improve its predictive performance over time as its experience improves
  • This process is done without explicitly being programmed to do so
  • Now since AI/ML concepts are possible to stand upon, we can make them powerful and economical through the usage of cloud technologies
  • Amazon Rekognition is a CNN based service, in which we can identify objects, people, texts in images, videos, and very accurate facial analysis
  • In this, it results in a confidence score on each label it produces, this is the same as a prediction for a regular ml model
  • AWS Rekognition even provides coordinates of an object detected in an image
  • OpenCV is a module in python that lets users access the webcam or any connected camera to show its content/snap a picture and so on
  • IAM is an AWS service in which we can create and assign power to a user
  • Amazon Lex is a service meant for building conversion based interfaces into an application via text and voice ip/op this internally uses NLP for communicating
  • NLP is a subset of machine learning, as in its the sub-field in ML which is solely focused on developing a computer’s ability to comprehend and respond to human language
  • Amazon Kendra is an intelligent search service, but in order to use it we must create an index and add the required data sources of the list ( it may be of such as S3, SharePoint, google drive, and so on~ )
  • Amazon CodeGuru is a service that helps us review and optimize our code for more proficient use, it even has a module in python called “codeguru_profiler_agent” for accessing the CodeGuru profiler
  • Amazon Comprehend is another NLP-based service that can be used to discover/obtain information in unstructured data. Its dome of identification consists of : objects, places, names, language, and so on~
  • If we wish to extract texts from handwriting and scanned documents we can use an ML-based AWS service called Amazon Textract
  • We use the Amazon Translate service if there’s a need to translate from one to another language
  • Amazon Personalize is the service that is meant for applications which are for recommendations like an e-commerce website
  • Amazon also provides a Fraudulent detection service called Fraud Detector, this manages and makes it easy to detect potential fraud activities online such as scalping, online payment fraud, fake accounts, and so on~

I’d like to thank Mr.Vimal Daga Sir, Preeti Mam, and the whole LinuxWorld Informatics Pvt Ltd team for organizing such a great session.

Thank you for the time~