Welcome to the Deep Dojo blog. A guide to machine learning on the Mac. If you’re interested in machine learning news and how it intersects with developing software for Apple hardware, you’re in the right place.
Today we find ourselves in the middle of a technological explosion. Machine learning has made its way out of the fringes of artificial intelligence and into state-of-the-art. The progress and adoption we’re seeing, particularly since 2015, is astonishing. It’s hard to wrap your head around the implications without some kind of analogy.
Bear with me for a moment…
Database-Enabled Disruption
You are a genie. With a snap of your fingers, relational database technology disappears from existence. Google, Facebook, Amazon and Twitter (which have a combined market cap of over $1.5 trillion) start their day tomorrow without being able to use database technology for their business.
What is left of these companies when this happens?
Along with the Internet, databases sit at the foundation of these companies. It enables them to do what they do.
We’re seeing the beginning of this enabling cycle for machine learning.
Just like it would have been hard to predict Google or Amazon when SQL was standardized in the early 90’s, we have no idea what the future holds for new companies enabled with machine learning. One thing we can be sure of is it will enable capabilities no one is currently able to do now.
Mac Lovers
If you’re serious about deep learning these days, you’re buying NVIDIA graphics cards and training your deep networks on an Ubuntu rig.
Can you do this kind of training on your Mac? Yes. Is it as fast as an NVIDIA-laden Linux box? No, but that shouldn’t stop you from experimenting on your Mac to see what all the hubbub is about.
There are WWDC talks. There are hardware accelerated math libraries.
If your network is already trained, there are even GPU-accelerated API for executing deep networks on iOS and tvOS. We’ll look into this and other developments for Apple hardware in future posts.