what up wonderful humans
!date
that's right. you're looking at a blurb. what's a blurb?
a short piece of writing that describes or summarizes something longer and more complicated
that's a meta-blurb. and for my purposes that was the third or fourth result on google.
i rented this book from the library called
because i read some reviews that it was good and in the beginning the author explains how dumb you can be and still pick up machine learning.
calling all dumbos.
just kidding he doesn't say dumb but what was important for me is he specifies that you don't need either the high level understanding of math or mad programming skillz to learn how to apply, use, or build machine learning...things.
this blurb is getting away from me but all seminal blog posts do.
i've always wanted to take a crack at machine learning but never really knew where to start. i'm deciding to put all 100% of my faith into this book. just kidding but i simply need to start, and not knowing things like calculus or not being able to explain to a random passerby what the word algorithms really means (because they always run away..how am i supposed to cement this knowledge if strangers won't take my word for granted), i feel like i should give it my attention.
what's with the blurb stuff? you wonderful people and you're wonderful questions.
i'm just starting to get back on the horse of studying again while entertaining a full time job. since getting into software support - all day everyday i spend solving tech issues. and that's hard. during the entirety of my last position, i barely programmed at all because after solving everyone else's problems for 8 to 12 hours a day, i didn't really have it in me to (create and) solve my own programming problems.
so the way that i am picking it back up (the learning horse, it's heavy) is a little bit at a time (the horse analogy falls apart here...and i'm there to pick up the pieces!). and since machine learning is super duper complicated i decided my readership (hi mom!) might appreciate some bite sized chunks to work off of. so tonight starts my series of blog posts to help me practice and hopefully for you to...i honestly don't know what everyone comes here for this is really all for my learning but if you learn something, aren't you the real winner?
first stop.
what is machine learning?
machine learning is a vein or area of programming that teaches computers things so that those computers can come up with solutions to their own problems.
why would we want computers to solve their own problems?
sometimes there is no real solution to some of our huge questions or sometimes, the way to answering those questions is far too arduous/costly/painstakingly BORING.
how machine learning works is we have a bunch of data and we feed it into a program that then can use that data to make predictions or find patterns in new data.
example time!
one of the simplest, and first commonly used machine learning programs was a spam filter for emails. you could write a spam filter yourself to look for certain words or phrases in your emails that would automatically detect what is junk and what is not however, you would constantly need to update this program to keep looking for new phrases and new words.
machine learning makes it's own rules based on previous data that is loaded into it's programming. what's more, certain email clients have a button that you can mark emails as spam to teach the filter what you believe to be spam. this is like adding a rule to the filter yourself and you have just taught your email robot army to detect a new kind of spam.
machine learning has grown up from spam filters and we see it in many places now. we see it in online shopping to classify and serve up similar or related items you might buy, teaching self-driving cars that yellow lights mean go faster (right?), and even my Instagram feed has come to understand that not only do i love hilarious cat and cool dancing videos, but that i REALLY love hilarious cool dancing cat videos.
here's an awesome list of other things machine learning is good for so i don't have to keep coming up with super duper creative examples on my own.
- making predictions (house prices, business revenue, the weather, stock prices)
- summarizing long documents
- designing a program to understand voice commands
- detecting credit card fraud
- reading handwritten digits
- designing intelligent bots for games
and the list goes on. there might even be a machine learning program that comes up with lists!
so to wrap up because this is so far from being a blurb at this point - machine learning is good for problems to which there might not be a clear solution. it can be used to help understand data better as well, finding patterns in large datasets and the machine learning programs themselves can be checked to see what has been learned. that's right, even we humans can learn from machine learning!
this was just the first 6 pages of the book. next, i'll get into the different ways that machines do the learning. until then..do whatever you do when you're waiting for me to write another blog post.
or whatever.
thanks for reading.