You would think that when it comes to making compositions and music, you’d need a really good ear and knowledge of the arts. Not so much with Amazon Web Service’s new AI (Artificial Intelligence) service focused on creating musical pieces with a keyboard! DeepComposer is the latest in a series of Machine Learning focused services that AWS has introduced since it’s announcement of DeepLens at Re:Invent 2017.

The new music based AI is a 32 key, 2 octave keyboard which will allow developers to familiarise themselves with using Generative AI. The simple application of Generative AI in DeepComposer will take short riffs and generate a full compositions.

The DeepComposer generative AI will be able to layer and generate songs based on pre-trained models or even user defined models. The pre-trained models are able to generate based on algorithms developed by training the AI with large musical data sets. The user defined models give users better control of the generative AI. Users will be able to define multiple parameters including the Architecture and Discriminator. The latter allows the AI to distinguish between the genres and determine the overall composition.

Being a machine learning model, DeepComposer is continually learning to identify music types. The AI will improve with time as it learns and generates more music based on the models and riffs. It will also be able to generate music which mimics a defined model. Amazon’s release touts, ” you have to train as a counterfeiting expert in order to become a great counterfeiter “.

DeepComposer isn’t just linked to the physical keyboard. It also has a digital keyboard interface which allows users to compose on the go. Using this approach, AWS is hoping that Generative AI models are made more approachable for those looking to explore their applications.

The new feature is currently available for preview on AWS at the DeepComposer website. Also on the website is a FAQ to address some of the questions that new users may have.