AWS and Microsoft have joined forces to announce Gluon, a new open source deep learning interface, giving developers accessibility to easily and quickly build machine learning models for their apps and other services.
Typically, the process of building a neural network can take as long as a week or more, as developers train network models to finely ingest large volumes of data. Deep learning engines like Apache MXNet, Microsoft Cognitive Toolkit, and TensorFlow have emerged to help optimize and speed the training process. However, the models are required to “learn” how to better parse out the datasets as they are provided, already a challenging prospect for even the more experienced developers. Besides being time-consuming, it’s also labor-intensive, often requiring developers to write long lines of complex code that are hard to modify, debug and reuse. While other deep learning tools could make model-building easier, this simplicity can come at the cost of slower training performance.
Gluon provides developers with a concise, easy-to-understand programming interface that enables them to instantaneously prototype and experiment with neural network models, and a training method that has minimal impact on the speed of the underlying engine. Developers can use the interface to create neural networks immediately, and to change their size and shape dynamically that makes them easier for adaptation. In addition, because Gluon brings together the training algorithm and the neural network model, developers can perform model training one step at a time. This means it is much easier to debug, update and reuse neural networks.
With Gluon, the programming interface hopes to remove prohibitive barriers for all developers besides those experienced in A.I. fields, and a chance to quickly prototype and experiment with neural network models without sacrificing performance.
“Today’s reality is that building and training machine learning models requires a great deal of heavy lifting and specialized expertise,” said Swami Sivasubramanian, VP of Amazon AI. “We created the Gluon interface so building neural networks and training models can be as easy as building an app.”
Available on GitHub, Gluon currently supports Apache MXNet. Am upcoming release will add support for the Microsoft Cognitive Toolkit, as well as other frameworks. Developers can build machine learning models using a Python API and a range of pre-built, neural network components.
Gluon is not the first time that AWS and Microsoft have collaborated around AI. Both companies are actively working on getting both the Amazon’s Alexa and Microsoft’s Cortana voice-command assistants to communicate with each other.. And last September, Amazon and Microsoft — along with Facebook, Google and IBM — announced the Partnership on AI to collaborate more on research and best practices in this newly emerging area.
This collaborative relationship between both companies reflects more promising developments in deep learning. “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,” said Eric Boyd, corporate vice president of AI and Research at Microsoft, in their press release.