Amazon Web Services Inc. (AWS), announced today a new machine-learning service at its summit, promising to simplify technology.
Machine learning (ML), which uses an algorithmic approach for analyzing historical data, is used to create predictive models and guide decision making.
These models can be used to filter spam, detect fraud, personalize content, retain customers, and market among other uses.
AWS announced Amazon Machine Learning. It stated that it wants to make the technology more accessible to the general public, much in the same way Big Data is becoming “democratized” through simplified usage so that anyone can use the technology, even non-specialists.
AWS stated that “Until now, very few developers were able to create applications with machine-learning capabilities because doing so required expertise data analysis, statistics, and machine learning.” Machine learning is a tedious, repetitive and error-prone process that involves manual tasks. These include computing summary statistics, performing data analyses, performing data analysis, and using machine learning algorithms to train models based on data. The model can then be evaluated and tuned, and then the model can be used to generate predictions. Amazon Machine Learning makes machine-learning accessible to all software developers. It abstracts away the complexity and automates these steps.
AWS stated that the fully managed and highly scalable service is built on the same technology internal developers use for more than 50 billion predictions per day.
Developers can access the service’s APIs (or its AWS Management console) and use visualization tools and wizards, to create ML models that work with data from stores like Amazon Simple Storage Service (Amazon S3) or Amazon Redshift (Amazon RDS).
“Amazon has a long history in machine learning. It powers product recommendations customers get on Amazon.com. It is also what makes Amazon Echo respond to your voice. It is also what allows us unload entire trucks of products and make them available for sale in as little as 30 seconds,” stated Jeff Bilger, a company executive. “We recognized early on that machine learning was only possible if it was made available to all Amazon developers. Amazon Machine Learning is the culmination of all the lessons we’ve learned while enabling thousands upon thousands of Amazon developers to quickly create models, test, and scale to power planet-scale predictive apps.
AWS stated that the ML service is available as a pay-as you-go service. This allows developers to start small projects and then scale up as required. Pricing information is available based on the different fees for data analysis, model building, and storage.