What is machine learning?

© Dmitrii Shironosov / 123RF.com

I won’t make that mistake a second time, you resolve, and learn from experience. If you translate the experiences into data, computers can do the same: They improve without a programmer telling them every step.

In May 2022, a nice person from Linux Magazine wrote to me and asked if I could write a short article on the question “What is machine learning?” could write as an introduction to the August issue. After a few short queries, I agreed, as you can see here. I try to implement the request more creatively than just chasing the detailed English Wikipedia article through the translation service DeepL and then shortening it. That would have been a first answer: Machine learning – DeepL is based on it – is an approach to cheaply generate templates for articles in German.

Now that we’ve known each other a few lines, allow me to ask you a somewhat personal question: As you read the opening lines, did you see a man or woman from the Linux Magazine editorial team in your mind’s eye? Second question: whoever it was, can you imagine why? Most of us compose images from our experiences as we read. Since you had no information, your brain used the experiences it had or thought it had with media in general or Linux Magazine in particular.

So you use experience to infer new, unfamiliar situations, and you do it all the time without it being particularly obvious. Of course, that doesn’t always work, but that’s often not a problem. Machine learning aims to give computers the ability to learn from experience and generalize. With computers, experience is…

[…]
Linux Magazine Online publishes all print articles that have appeared in Linux Magazine since 2001. This means that you have access to a high-quality archive including the articles from the current issue online. Most of the more than 4,000 articles are accessible free of charge, only articles (as PDF) from the ten most recent Linux magazines require a small fee.

#machine #learning

Leave a Comment

Your email address will not be published.