From now on, Microsoft and Facebook will be working together to provide an open source AI resource repository, which should help the community develop more and more advanced AIs.
Called the Open Neural Network Exchange, or ONNX, this platform will allow researchers and developers to share algorithms between AI engines at various stages of development. From them, everyone can learn.
This is pretty much the first time when AI research gets such an opportunity, and it’s certainly going to have success, if we only look at how well other repositories have fared over the years and how much of an important role they’ve ended up having.
The two companies explain that ONNX will be supported by Cognitive Toolkit, Caffe2, and PyTorch, which are all frameworks that provide interfaces that make it easier for developers to build and run computation graphs that represent neural networks. “ONNX is the first step toward an open ecosystem where AI developers can easily move between state-of-the-art tools and choose the combination that is best for them,” said Facebook Director of Applied Machine Learning, Joaquin Quinonera Candela, in a blog post.
ONNX will bring several key benefits to the table. First off, they’re going to allow developers to move easily between frameworks and use the best tool for the task they’re working on. The platform will also allow hardware vendors and others with optimizations for improving the performance of neural networks impact multiple frameworks by directly targeting the ONNX representation.
Microsoft and Facebook seem to be sharing the same vision of making AI more present in our world and easily accessible to everyone. This makes ONNX a very valuable tool to everyone working in AI research, as well as consumers who will likely see more implementation of artificial intelligence in the tools they use, from apps on their phones and computers, to self-driving cars.
The hope throughout the community is that other companies will also join in the open-source platform and start sharing what they’re working on, such as Apple, a company that’s known for its secrecy.