It’s no secret that machine learning is kind of a big deal. Luckily, no matter where you are in your career (or the world), you can dive head first into the crazy cool world of machine learning.
If you’re looking to not only transform your career and yourself, but also to position yourself at cutting edge of cool, then keep reading my friend, because we’re about to turn some lead into gold.
If you’re like me and don’t happen to come from a computer science or math background then getting started with machine learning might feel pretty overwhelming.
You might be wondering how to even begin.
It’s possible that you’ve already done a quick google search and come up with academic papers full of eye-wateringly long formulas and arcane terms, forums filled with developers passionately debating extremely technical details, or fluff pieces that are full of buzzwords without actually telling you anything useful.
When it comes to non-tech people getting involved in machine learning, it can feel like a members-only party. But it doesn’t need to be exclusive, and contrary to popular belief, machine learning is totally learnable, even by someone who knows nothing about computers!
As with any new endeavor, tackling machine learning is a combination of taking a few simple steps and lots of elbow grease 💪💪 Most of all though, it’s about just starting.
Getting started with machine learning is no different than starting any other project. Here’s what you’re going to need:
A powerful-enough computer that is machine learning ready
You can’t make machines learn without a machine, regardless of whatever sub-field you end up getting into, you’re going to need a computer to get the job done. No joke.
For simpler algorithms a basic computer that you have administrator privileges for (aka you can install software on) will work. If you’re looking to start running ASAP, then just install the software which we talk about below 👇 and you’ll be able to get going.
If you’re interested in trying out some of the more complicated algorithms like neural nets or generative adversarial networks, you’re going to need a computer with higher processing power.
Some algorithms are really intensive and if your computer isn’t powerful enough, a project that should take an hour might take days or even weeks to complete.
Ideally, this means a computer that has a minimum of 8gb of RAM, a powerful graphics card (like the ones in NVIDIA’s GEforce 10 series), at least an Intel i5 processor, and 1tb of hard drive space.
General rule of thumb = the faster your computer the more fun you’ll have.
Once you have your computer, you’ll want to install Anaconda and then install some of the major machine learning libraries like Tensorflow.
Understanding of key concepts
Artificial intelligence is a pretty big (and frequently misunderstood) discipline that incorporates a number of sub-fields, including machine learning.
To understand what AI is and isn’t, as well as its potential it’s really important to have a broad understanding of the key concepts. If you get even cocktail party levels of knowledge about key topics such as machine learning, deep learning, and neural nets, you’ll already be ahead of many people.
Plus, you’ll sound really smart.
There’s no math or programming skills required to learn the basics of how machine learning is being used to identify the difference between pugs or loaves of bread or how natural language processing lets you talk to your phone, which means you can start reading up on these topics today 🙌
Overview of the fundamental algorithms
Machines learn by following algorithms. Fundamentally, an algorithm is a list of rules to follow in order to solve a problem. They tell a computer what to do and when.
There are many, many different algorithms out there that do all manner of things, but the basics of machine learning are built on a few tried-and-true ones.
As with most things in life, when it comes to algorithms it’s best to start out simple.
Some coding skills
If algorithms tell a computer what to do, you communicate this message to your computer through programming (aka coding).
Before you dive into actually implementing algorithms, you’re going to want to decide which programming language you will use.
When it comes to AI and data science, there are two major players – Python and R. You’ll want to decide which of these languages to focus on first. Once you know one programming language, it’s easier and faster to learn additional ones so you can always switch later on.
If you’re new to the scene then Python is most likely your best bet as it’s one of the fastest growing programming languages and is used in many different scenarios beyond machine learning.
In short, learning Python for machine learning means you’ll also be able to do all sorts of other cool (and in demand) things.
Basic math skills
Pretty much everything in the world can be explained through math. It turns out that math is actually really, really cool.
I know, I know. It’s likely that you’re still suffering from math related traumas. I get it. But now that I’m an adult (🤔) and not forced to sit through math class I’ve discovered the incredible secret that they never teach you in school – math is awesome.
So if the thought of numbers makes you panic you’ll probably want to brush up on some basic math. Feel free to go as basic as you want – no one has to know.
Machine learning is based on linear algebra, statistics, probabilities and some calculus. You won’t need these to get started but as you get deeper into machine learning you’ll want to start picking up some of the fundamentals.
Luckily, you won’t need to learn mathematician levels of math, just enough to get your computer to do what you want.
Data that’s ready to go
Data is to machine learning as gas is to your car. The outrageously massive volumes of data that humans create every minute is part of the reason machine learning has been able to make such incredible advances lately – there is just so much data!
When you’re starting a machine learning project you’re going to need data. Luckily, there are a number of really high quality datasets out there that are free (!). Here’s a good place to start.
Once you have a dataset that you want to use, odds are it’s going to need to be cleaned and pre-processed. This involves tasks like filling in any blanks in the data, making sure the numbers are all on the same scale, and converting text data into numerical data. This can take more time than you think so schedule accordingly.
Willingness to make mistakes and try new things
Oh yeah, and get ready to make some mistakes. But in a fun way.
Being a beginner at something can be really frustrating but it’s totally worth it. It’s pretty well established that approaching a project with a beginner’s mindset is where the magic is.