We’re working to bring the power of neural nets to every application, using new technology invented and built in-house, to make applications that weren’t possible, possible. There’s a large gap between the capabilities neural networks show in research and the practical challenges in actually getting them to run on the platforms where most applications run. Making these algorithms work in your app requires fast enough hardware paired with precisely tuned software compatible with your platform and language. Efficient plus compatible plus portable is a huge challenge—we can help.
Today’s CPUs and GPUs are powerful and efficient enough for many intelligent applications. Demand from gaming and graphics has pulled unprecedented amounts of computing power into devices all around us, enough to make our technology a whole lot smarter. That’s about where the good news stops. The lack of portable, developer-friendly tools prevents most organizations from realizing the power of deep learning for their business. We’ve talked to many companies across a range of industries—over and over we hear the same thing: It’s too hard to learn the algorithms and write the low-level code, there are too few developers with the right low-level experience. The hardware is good enough, but building the software is too risky and too costly.
A year ago we saw a way to solve the compatibility and portability problems all at once, for all platforms, using a new software approach. It’s required rethinking how we implement the algorithms, and it’s been a challenge to engineer—but the payoff is worth it. Bold claims require bold evidence—over the coming months we will be sharing what we’ve built, much of it as open source software. If you’re struggling to deploy intelligent applications based neural nets to cloud, embedded, mobile or desktop targets—we’d love to hear from you.