The senior vice president of Amazon Web Services, Andy Jassy announced the new machine learning platform at the AWS Summit in San Francisco. The new machine learning platform is designed to give developers, without a machine learning background, the tools to build data-driven smart apps that can monitor what is happening in real time. The machine can also predict whats going to happen in the future.
Many companies have been using AWS tools for machine learning tasks but its tough work. AWS’s objective was to create a service that did not require any special skills. Amazon has been experimenting on machine learning.
Developers are constantly exposed to data and need to interact with data in real time. This would help in fraudulent transactions in e-commerce site. AWS has the ability to recognize that there is a lot of technical expertise involved like statistical analysis, model building, data cleansing etc.
Amazon is now looking to make the process simple and effective. This will allow developers to build a model and process billions of predictions in real time.
There are three basic steps:
- Build your model in Amazon S3 or Redshift.
- Validate and optimize your model.
- Make batch predictions with trained models.
Amazon conducted a few experiments to test the platform. First, two developers tried to predict customer gender; surprisingly those 2 had increased the gender accuracy from 65 to 95 percent. The 2 developers did not had any specific expertise but were still able to achieve 95 percent. However, Mr Jassy claims that it took him only 20 minutes to reach that level of accuracy.