More than a month ago, Microsoft and Amazon teamed up to integrate their voice assistants. Today, the companies have announced their collaboration in the formation of Gluon, a new deep learning library which allows developers to build sophisticated machine learning models with relative ease.
The Gluon interface will help developers to build machine learning models by providing them with a Python API, along with prebuilt neural network components. Additionally, it will allow them to debug and update neural networks much more smoothly. Currently, the deep learning library only works with Apache MXNet, Microsoft has noted that support for Microsoft Cognitive Toolkit (CNTK) will be available soon.
The problem with the construction of neural networks seems to lie in maintaining balance between model building and training performance. Note that deep learning engines such as Apache MXNet, Microsoft Cognitive Toolkit and TensorFlow do optimize the training processes to some extent, but often require a lot of time and complex coding, on the developer"s part. The Gluon interface will provide developers the ability to experiment with various neural network models, along with a training method which has barely any impact upon the performance of the underlying engine.
CVP of Microsoft AI and Research, Eric Boyd, believes that the Gluon Interface will provide developers with a ""freedom of choice"". To that effect, he noted:
We believe it is important for the industry to work together and pool resources to build technology that benefits the broader community. This is why Microsoft has collaborated with AWS to create the Gluon interface and enable an open AI ecosystem where developers have freedom of choice. Machine learning has the ability to transform the way we work, interact and communicate. To make this happen we need to put the right tools in the right hands, and the Gluon interface is a step in this direction.
Only time will tell how useful the interface will prove to be to the machine learning community, and whether it will be as efficient as is being claimed. In the meantime, you can get started with Gluon on Github right away.