Amazon Aurora Machine Learning is a proprietary technology of Amazon that enables a native SQL user to integrate Machine Learning-based predictions to an application without knowing or understanding any machine learning algorithms.
Machine learning heavily relies on datasets for it to work. You can say that data is the oil that keeps the engine of machine learning running. There is a massive amount of data generated every day. To give you an idea, according to this article, “By 2020, it’s estimated that for every person on earth, 1.7 MB of data will be created every second.”
Almost 2 MB of data may not sound a lot to you but you should not forget that the current population of earth is 7.8 Billion! Crazy right? Now, I’ll let you do the math. Machine learning helps businesses to get better insights into their data to make better decisions. In most cases, structured data gets stored on a database. Today, to use machine learning with a relational database, you need to build a custom application to read the data from the database and process it using your model.
Much like a middleman between your database and application:
Building this kind of application for machine learning can be tedious and requires the experience and knowledge of a machine learning expert. Amazon makes it easy for you to apply and experience the benefit of machine learning even if you still don’t have any experience with it. As of this writing, Aurora is natively integrated with two machine learning services:
- Amazon Sagemaker – is a fully managed service that provides every developer and data scientist with the ability to build, train, and deploy machine learning (ML) models quickly
- Amazon Comprehend – is a natural language processing (NLP) service that uses machine learning to find insights and relationships in a text.
With these services, the middleman which reads and processes the data from the database is eliminated. You no longer have to think about the underlying process on the back-end because SageMaker and Comprehend is fully managed by Amazon. For example, you can use Amazon Comprehend through SQL queries in Amazon Aurora to scrape information about whether your product has a positive or negative review.
Since Amazon Aurora has a direct connection to Comprehend and Sagemaker on the back-end, it provides a low-latency connection that is suitable for real-time requests using ML-based predictions on large amounts of data.
Common Use cases:
- Fraud detection – detecting suspicious activity and patterns on transactions to separate legitimate operations from shady ones.
- Ad targeting – sending specific ads to relevant users on social media platforms.
- Product recommendations – providing recommendations for product listings based on what others are buying with the same item.
- Sentiment Analysis – processing of text data using NLP (Natural Language Processing) to identify a customer’s sentiment towards a certain product.
- MySQL 5.7
- PostgreSQL 10
- PostgreSQL 11
You are not charged for the integration of Aurora with AWS machine learning services. You only pay for the resources you use.