I probably don't need to tell you that people love throwing the terms "machine learning" around these days. Heck, we're at a point now where some percentage of them aren't just saying it to sound smart and know they actually know what it means. (I'm not one of them) 😈 (Is that an evil cat Emoji? Yeah!)
A big problem with the actual machine learning algorithms that are real things that are actually doing computations on computers is how black boxy they are now. A lot of it is one algorithm build on another and another and modern PCs just have the firepower to run all of them.
Now you have a new problem. What are they doing to your data back there!?! And if they are screwing it up, when would you know? In 2 weeks when all the calculations are done? Or when someone does a follow-up to your study and the prose has a condescending tone (that you probably imagined...😈...)
DALEX is an attempt to figure this out. It is a project by a bunch of bored mathematicians called MI^2 (this is pronounced "Am I square").
You can read about this project here.
Shoutout to Dr. Norris (someone who does fall in the group of people who knows what machine learning is and how to use it) for this cool link!
No comments:
Post a Comment