You gotta love your grandparents. They give great advice, have answers to almost every question, are always up for an adventure, and love their grandchildren unconditionally. Grandparents are awesome — except when you’re trying to show them how FaceTime on an iPhone works.
Grandparents are slowly coming to terms with technology, but explaining it to them can be pretty difficult. For example, my grandma calls Mozilla Firefox the Godzilla flame wolf… and I’m not correcting her.
So how do you explain machine learning in a way that a grandmother who does not have a background/experience in STEM can understand? In order to do that, I won’t get too technical, but I will get in-depth into the topic and explain it in simple terms. I’ll cover the most important parts of machine learning — the big pieces everyone should know.
To make sure I don’t get too technical, I’ll be referring to you, the reader, as grandma from time to time. It’s a way to remind myself that I need to explain this as simply as possible and not get ahead of myself.
So, Granny, what is machine learning? Machine learning is a field of study that gives computers the ability to learn without being explicitly programmed. It is a very exciting and growing field that’s making computers more similar to humans. It is different from traditional computer science approaches. In regular computing, instructions are given to the computer to calculate and solve. Computers are really dumb and have to be told exactly what to do and how to do it. With machine learning, though, we can give a computer lots of data to analyze. It can then train on that data to produce output values that fall within a specific range.
For example, you know how you post pictures on Facebook, and it recommends people you should tag because they too might be in the photo? Oh, you don’t? Okay, how about when you’re watching Netflix and it recommends shows or movies you might like? That’s a little taste of machine learning.
When cars drive themselves, machine learning will play an important role. Cars will collect lots of data to learn how to drive better and safer. Hopefully, this is making a lot of sense because we haven’t got into the best parts of machine learning. One thing is for sure, though: it will play a huge role in our lives going forward.
What Machine Learning Is Not
Machine learning is not about robots coming out to destroy humans like you see in the movies. “Terminator” is the first thing people think of when they hear the term artificial intelligence or AI. That’s another thing I want to talk about. Machine learning is not artificial intelligence, but a subset of AI. This field has been around for quite some time, with the roots going back to the late 1950s. During that time period, IBM’s Arthur L. Samuel created the first machine learning application, which played chess.
Another buzzword you’ve probably heard is deep learning. Deep learning has been around just as long as machine learning, but it wasn’t until the 1980s that the field gained traction. Eventually, big companies like Facebook, Google, and Microsoft started investing heavily in the technology. The result has been a revolution for AI. Things like Google Translate or Apple’s Siri are examples of the power of this technology.
I won’t get into what AI or deep learning is as machine learning is already a big subject to cover by itself. Just know there is no threat with machine learning — maybe with AI, though, if it gets out of hand.
How to Get Machines to Learn
So you might be thinking, how exactly do we get these machines to learn? How does a computer collect all this information and make sense of it? Well, I can tell you that there’s a lot of math and algorithms involved to help produce the desired results. Alright, Grandma, I’m gonna break it down a simple as I can for you. But at the same time, I’ll explain in detail what’s under the hood of a learning machine.
The math that comes with machine learning
Math was never my favorite subject, but we’ve all come across linear algebra. Linear algebra is a field of mathematics that is universally agreed to be a prerequisite to a deeper understanding of machine learning. Linear algebra is a large field with many esoteric theories and findings. But the field’s basic tools and notations are helpful for machine learning practitioners. With a solid foundation of what linear algebra is, it is possible to focus on just the good or relevant parts.
Math is important in this field because you need to know which algorithms to include when considering accuracy, training time, and a bunch of other stuff. The math helps us find a way to help the machine learn in the best way possible. Other than linear algebra, a scientist/engineer will also need to know the mathematical concepts of Calculus, algorithms, probability theory, and statistics. Python is the most used programming language in this field as well. A beginner doesn't actually need a whole lot of math to get started in this field.
Think of a brain
The world has a lot of information, and our brain takes it all in to form our view of reality. A computer has to be able to do the same, and that’s where a neural network comes into play.
Neural networks are the most popular way to get a computer to mimic a human brain. Your brain is made up of approximately 100 billion nerve cells, which are called neurons. Our brains are really good at solving problems, and each neuron is responsible for solving a tiny part of that big problem. They have this cool ability to collect and send signals. Think of it like a network with a bunch of wires.
So computers’ neural networks are inspired by the brain. Now, you may be wondering how the neurons are linked together. A neuron takes in inputs and produces outputs. The input nodes (input layer) provide information from the outside world to the network. Similar to how your eyes see and collect information and send it to the brain.
The output nodes (output layer) are the ones responsible for communicating that information back to the outside world. Let's say that the network below is trained to recognize digits. So a number in the input layer will go through the hidden layer. Then, it will come out the output layer as the number being recognized. The neurons in the hidden layer are going to communicate the information they got from each other. They’ll use this information to piece together what number they think is being passed along. Each layer influences the next.
Things get even cooler when you’re training the computer to network for other things, like audio recognition. The computer can learn how to parse speech, break down audio, and pick out distinct sounds. These sounds are combined to make certain syllables, words, phrases, and etc. This is how you want to think of it when building your own network:
- Convolutional network— for image recognition
- Long short-term memory network- good for speech recognition
There are other ways for a machine to learn, including supervised learning, unsupervised learning, and reinforcement learning. I won’t be covering those topics, but those are the three methods that are often employed. To put it simply, Grandma, a neural network lets the computer take information, break it down into pieces it can understand, and then outputs the closest outcome it can.
The challenges and limitations
As awesome as machine learning is, there are limitations to it. I’ll talk about the biggest ones I think need to be overcome for this technology to move forward. So machine learning algorithms require massive stores of training data. Labeling that data is a tedious process. You need to make sure that the data being fed into the machine is labeled. If not, it’s not going to get smart over time. An algorithm can only develop the ability to make decisions and behave in a way that is consistent with the environment it’s required to navigate.
Another problem is machines can’t explain themselves. That can be difficult when you want to know why the machine made a particular decision.
Lastly, and most importantly, is avoiding bias. Transparency is important and unbiased decision-making builds trust. For example, facial recognition plays a large part in social media and law enforcement. But biases in the data sets provided by facial recognition can lead to inexact outcomes. If bias finds their way into an algorithm and data sets and the training data is not neutral, the outcomes will amplify the discrimination and bias that lies in those data sets.
The Future Is Machine
The future of machine learning is unstoppable, and it’s the fundamental building block for artificial intelligence (AI). Today, it already plays a role in our lives. If you use Spotify to listen to music, you will see it creates daily mixes for you based on what you’re listening to. Amazon learns and teaches itself how to get products that might be of interest to you based on your buying habits. Virtual assistance like Amazon’s Alexa, Apple’s Siri, or Microsoft’s Cortona use machine learning to help understand the language humans use when they interact with them.
Businesses are obsessed with this technology because it can automate tasks normally done by humans. Many companies use chatbots and service bots in their customer service departments. These bots are learning to respond in an intelligent and helpful way to customers.
My favorite place to see machine learning being used is in autonomous cars and trucks. Vehicles need to be able to understand obstacles in the road and how to respond to them, including stop signs, snowstorms, a ball in the street, or another vehicle. The more data they collect, the more human-like they start to act. For example, they’ll know a stop sign covered with snow is still a stop sign.
So there you have it, Grandma. I know I lost you somewhere, but hopefully, you have a better understanding of what machine learning is. I see machine learning as a tool that can continue to make our lives easier. People are continuing to come up with helpful ways to use machine learning, and they’re disrupting industries by doing so. I can only imagine where we will be when this technology leads to real AI.