AWS Machine Learning comprises of a rich set of tools that Amazon offers to help developers integrate machine learning models into their applications. AWS offers pre-trained models for use cases including computer vision, recommendation engines, and language translation.
Amazon currently offers 15 machine learning services on its platform. In this article, we will take a look at each one of them and understand the type of problem they try to solve for their customers.
Amazon SageMaker helps you to take your machine learning models from concept to production in a fraction of time compared to traditional code-based approaches. Sagemaker is a managed service with a suite of tools required to build, train, and deploy machine learning models. From labeling your data to optimizing the algorithms, SageMaker is a complete solution to create and deploy machine learning models.
SageMaker also has an autopilot option that will automatically process the data and run it through multiple algorithms. This helps developers to find the best algorithm for their model without manually training and testing their models. Sagemaker also comes with an integrated IDE and a sharable jupyter notebook that you can use to collaborate with your team.
Amazon CodeGuru helps you to automate code reviews and performance optimizations for your application. CodeGuru can spot problems like race conditions, resource leaks, and wasted CPU cycles, helping you produce higher quality code by providing specific recommendations based on code context.
Codeguru’s algorithms are trained with codebases from Amazon’s projects. Right now, CodeGuru supports only Java applications, but we can expect the functionality extended to other languages in the near future.
Amazon Comprehend is a Natural Language Processing service from Amazon that uses machine learning to find valuable insights from textual data. Comprehend can work with unstructured data like product reviews, customer emails, twitter tweets, etc. and help you draw conclusions like overall audience sentiment.
Comprehend is also a fully managed service, meaning you can use pre-trained models to work with your data. Comprehend also has an additional service called Amazon Comprehend Medical that lets you work with medical documents to analyze medical conditions and dosages.
Amazon Forecast is used to build time-series prediction models using your existing datasets. Forecast can be applied for use cases including predicting future business expenses, stock price prediction, and resource planning for organizations based on customer demand.
Forecast is also customizable and lets you build custom models on top of Amazon’s existing deep learning models. Like most machine learning tools in AWS, Forecast is also fully managed and can scale according to your business needs.
Amazon Fraud Detector is another fully managed service that helps you detect fake registrations and fraudulent transactions. Fraud detector can identify potentially fraudulent accounts and helps you set up additional verification for those flagged accounts.
A fraud detector needs an existing dataset of labeled fraudulent transactions in order to train and understand the pattern of your customer behavior. This is used to prevent further fraudulent transactions. You can also configure custom authentication rules for guest logins and product trials.
Amazon Kendra is an AI-powered enterprise search engine that will help you deliver highly accurate search results based on customer queries. Kendra can be used to power search engines in your products that help customers find the exact information they are looking for.
Kendra can be used to help customers find answers to specific problems while using your product without the need for additional customer support. Kendra also supports natural language questions, delivering an even smoother experience for your customers.
Amazon Lex lets you build conversational interfaces into your products. Lex offers Natural Language Understanding (NLU) models that can understand conversational input from users and perform the right actions.
Lex can be used as a replacement for manual customer support to help filter the usual queries, answering them automatically. Lex is also a fully managed service that scales automatically and employs a pay-as-you-use model.
Amazon Personalize lets you create custom recommendations for your customers based on their usage patterns. While traditional recommendation engines can be used only to recommend products, Personalize lets you literally personalize every step of your customer’s user experience.
Personalize is a great tool to build product recommendations, custom search results based on queries and employ targeted marketing promotions.
Amazon Polly helps you build speech-enabled products for your customers. Polly provides lifelike voice outputs across a variety of languages including Chinese, Korean and Japanese.
Polly is powered by deep learning algorithms that mimic a conversational style interface that can be used in narrations, telephony applications, etc.
Amazon Rekognition is a computer vision solution from AWS that helps developers to build applications that can recognize objects from images and videos. In addition to automatic object recognition, you can customize Rekognition to pick specific objects and scenes based on your business requirements.
Rekognition can be employed in use cases like identifying manufacturing defects in products, spotting unauthorized personnel in an organization, scanning for inappropriate content in movies, etc. Rekognition can also be used to analyze player movements in games for post-game analysis.
Amazon Textract enables you to read data from scanned documents. The usual approach to digitizing paper documents is using manual data entry or OCRs with custom configurations. Textract makes it easier by automatically applying rules to documents and extracting valuable data along with components like forms and images within the document.
Textract is useful for processing loan applications, medical claims, etc. In addition to extracting data, they can be optimized for search using Textract. Documents that usually take months to process using manual methods can be processed in hours using AWS Textract.
Amazon Transcribe lets you build speech to text services in your application. Transcribe is useful in building medical transcription services, streaming audio, generating subtitles for video recordings, etc.
Transcribe can also be used to convert customer calls into text and analyze them for improved customer service. Cataloging audio archives is another use case for AWS transcribe.
Amazon Translate is a machine learning service similar to google translate. Translate can work with a variety of languages with high accuracy, enabling businesses to customize their language based on the audience demographic.
Translate is also designed to be more natural-sounding to customers since the context of the sentence is also taken into account. Translate is also highly customizable to help improve the accuracy of translation when working with your brand names and unique words related to your business.
Amazon DeepLens is a video camera with in-built deep learning capabilities that helps you to build and test computer vision models in real-time. DeepLens is fully programmable can be used to test models like live object recognition, classifying birds/ animals, face detection, etc.
DeepLens is designed for developers getting started in machine learning to get a grasp of how their models will work in the real world. DeepLens is also integrated with the AWS ecosystem and can be used with other AWS services like Lambda and Rekognition to extend its capabilities.
If you are a fan of self-driving cars, AWS DeepRacer is a small autonomous race car designed by AWS that runs using machine learning. DeepRacer helps you test your reinforcement learning models using a physical track.
You can build reinforcement learning models using AWS SageMaker and test them instantly using DeepRacer. Amazon also offers an opportunity to connect and compete with fellow enthusiasts by building virtual private race tracks.
With a solution to almost every machine learning problem, Amazon Machine Learning offers a rich set of tools for machine learning engineers to work with. Amazon also adds new services every few months based on new use cases, making it one of the most dependable platforms for engineers to build AI solutions for their customers.
Loved this article? Join my Newsletter and get a summary of my articles and videos every Monday.