Talking to people about machine learning is tricky.
I have learned this the hard way. Sometimes it feels like they don’t want to understand you.
In this article I am going to show you how you can talk about machine learning with normal people. Like your boss, other colleges, or your friends.
You just need to use a different language when you talk with non experts.
Never talk about algorithms and models
In 1959 the author Arthur Samuel coined machine learning as:
Field of study that gives computers the ability to learn without being explicitly programmedArthur Samuel
People understand, that if you want to solve a problem you need to explicitly program it. It has worked very well until now and it is what people know.
Without being explicitly programmed. This is where machine learning is totally different to typical problem solving.
That is why you should never talk about algorithms and models with people.
Here are some bad examples:
“This machine learning algorithm solves this or that…”
“This model tells you if this or that is good or bad…”
Algorithms & models. These two keywords tell people you explicitly coded a problem. Know matter what you say.
It is hardcoded in their minds.
So Andreas, what is the best way then?
Use learning, self-learning and patterns
Here’s what you can do:
Instead of algorithms use learning or self learning methods to describe how the solution works.
Try to avoid the term model. People associate with it that you specifically modelled something.
Talk about learning and self-learning. Talk about patterns and combine it with “the system”.
“This system learns how to solve this or that”
“This system self-learns how to solve this or that”
Or like that:
“This system has learned a pattern to tell you if this or that is good or bad…”
“This system has self-learned a pattern to tell you if this or that is good or bad…”
See what a difference this makes?
With this wording no one will think that you explicitly coded the problem.
Another way to solve this problem is by using the hammer method. The AI hammer.
Use AI even if it is wrong
What some people do, is to use Artificial intelligence instead of machine learning.
“This AI solves this or that…”
“This AI tells you if this or that is good or bad…”
From an experts view this behaviour is questionable. To use AI for some solution that is not an AI solution is, määäh.
But people know AI from movies and understand that this is some next level shit.
What is your recipe?
Do you have a special trick you use when talking to people about machine learning?
Tell me in the comments section below.
If you liked this article please share it with your friends on social media. That would help me a lot! Thanks
Wait, there’s more!
Liked this one? Checkout these articles I wrote earlier:
— 5 Scientific Working Tips For Data Scientists
Tips that help you to be recognised as a professional
— Do You Need A PhD To Be A Data Scientist?
The difference between research and applying
— 7 Super Simple Steps From Idea To Successful Data Science Project
Complete guide on how to do a data science project
— The Brutally Honest Truth About Learning Big Data the Right Way
Things you need to understand that makes starting with big data easy
Liked them too?
I write two to three articles a week. You can subscribe to my mailing list and I will send the link to your inbox when i hit publish.
Title Image by Cole Hutson on Unsplash