How Machine Learning Works: Explained Simply (Part 1)
- Angie Okhupe
- Feb 14
- 5 min read
A lot of people use Artificial Intelligence (AI) and Machine Learning (ML) interchangeably—but are they really the same thing? What do you think? Take this quick poll before reading further! 👇
Do You Know the Difference Between AI and Machine Learning? 🤖🧠
Yes, I totally get it!
Sort of… aren’t they the same thing?
Nope, I thought they were interchangeable!
Let's talk about driving and machine learning
I’ll be honest—driving stresses me out. But as an adult, it’s one of those things you kinda have to do. So, despite the overwhelming number of rules, the constant need to stay alert, and my deep fear of parallel parking, I had to learn.
At first, I wasn’t exactly a natural behind the wheel. I’d stall the car at stoplights, freeze up at four-way intersections, and reverse parking? Let’s just say, I often leave the car on the driveway and ask for help to back it into the garage. Parking, isn’t my strong suit!
But let’s focus on the actual driving part. The more I practiced, the better I got. I started to notice patterns—when to smoothly merge into traffic, when to ease off the gas before a sharp turn, and how to predict what other drivers might do. Over time, driving stopped feeling so overwhelming and started to become… almost automatic.
This a lot like how machines learn in Machine Learning (ML).
Before We Get Into the Nitty-Gritty… AI🧠 vsMachine Learning🤖?
Before we get into the nitty-gritty of how machine learning works, I'd like to clear up a common question: What’s the difference between AI and machine learning?
Here’s the deal:
Artificial Intelligence (AI) is the big idea—the broad concept of machines performing tasks that typically require human intelligence. This includes understanding language, recognizing faces, making decisions, and even playing chess at a world-champion level. AI can range from basic systems that follow fixed rules, like chatbots with predefined responses, to advanced systems that learn and adapt over time.
Machine Learning (ML) is a subset of AI—it’s one of the key ways AI systems become “smart.” Instead of following a strict set of rules, ML allows computers to learn from data, find patterns, and improve over time without needing explicit programming for every scenario.
Think of it this way:
AI is the big umbrella ☂️—it includes any technology that mimics human intelligence. ML is a tool inside that umbrella—a way for AI to learn from experience instead of following a rigid rulebook. AI isn’t just about machine learning—it includes other smart technologies, too such as:
Rule-Based Systems 🤖: Like Google Maps saying, “Turn left in 200 feet!” It follows predefined rules but doesn’t "learn" from experience.
Natural Language Processing (NLP) 🗣️: Your car’s voice assistant understanding, “Find coffee near me.”
Computer Vision 👀: A self-driving car recognizing a stop sign… or a squirrel darting across the road.
Robotics: Drone dodging traffic to deliver your burrito.
Decision-Making Systems 🧠: Google Maps rerouting you around an unexpected traffic jam.
While ML is the star learner in the AI toolbox, it’s not the only tool. Now that we’ve got that sorted out, let’s dive into how ML actually works!🚀
How Does Machine Learning Work? 🚗💡
Now we understand that AI is the big concept and machine learning (ML) is one of its most powerful tools. But, let’s get into the fun part—how ML actually works. And surprise! It’s a lot like learning to drive.
When I was learning to drive, I didn’t just memorize a rulebook and magically know how to navigate the road. I practiced, learned from mistakes, recognized patterns, and got better over time. That’s exactly how machine learning works too. Let’s break it down step by step... 🚗💨
Step 1: Practice, Practice, Practice 🏁
When I first started driving, I wasn’t thrown onto a busy highway. Nope—I started in empty parking lots, quiet streets, and eventually, more challenging roads. Of course, I made mistakes. I stalled at stop signs, turned too wide, and maybe hit a curb or two. But with each mistake, I learned something new. This is called training in machine learning. Just like I practiced driving with real-world experience, an ML model practices with data to recognize patterns and improve its accuracy.
Step 2: Recognizing Patterns 🔍
The more I drove, the more I started noticing patterns:
🚦 Yellow light? Slow down.
🌧️ Raining? Brake earlier and drive more carefully.
🛣️ Merging onto the highway? Match the speed of traffic.
Instead of memorizing every single driving rule, I started to recognize patterns and make decisions based on experience. Machines do the same thing. Instead of being explicitly programmed with a billion rules, they find patterns in data and use them to make predictions. For example: a spam filter doesn’t need someone to manually program every possible spam keyword. Instead, it learns from patterns—if an email has lots of capital letters, suspicious links, or phrases like “Congratulations! You’ve won a free iPhone!”, it flags it as spam.
Step 3: Building My Driving Skills (The Model) 🚗
After enough practice, I built my own driving instincts. I didn’t have to think too hard about when to brake or how to switch lanes—it became second nature. With ML, the machine builds a model—a set of learned patterns and strategies that help it make decisions. This is how a recommendation system (like Spotify) builds a model of your music taste. Instead of guessing randomly, it learns what you like and suggests songs that match your vibe - for me, its always an afrobeat anthem!
Step 4: Refining the Model 🔄
Even after I got my license, I kept improving. I learned to anticipate other drivers’ behavior, handle new roads, and adapt to surprises—like when it snows in Texas or a deer jumps in the middle of the road in upstate New York. ML models don’t stop learning after training—they keep improving by analyzing new data and adjusting their predictions. This is what makes ML a unique tool becasus like humans, it now has the capacity to improve!
Machine learning might sound complicated, but at its core, it’s just about teaching computers to learn from examples in a smart way. Just like learning to drive, machine learning isn’t about memorizing rules—it’s about practicing, recognizing patterns, building instincts, and improving over time. So, the next time you see a Netflix recommendation that’s absolutely spot-on or your phone automatically corrects your texts, just remember—it’s all machine learning. And the best part? The possibilities are endless!
Curious about the different ways machines learn? In part 2, I dive deeper into the three types of machine learning—supervised, unsupervised, and reinforcement learning—to help you better understand how these systems work and shape the world around us.
Bonus: Fun Facts About Machine Learning 🚀🤖
The First ML Program Was Created in 1952
Long before Netflix recommendations and self-driving cars, a scientist named Arthur Samuel developed the first ML program—a checkers-playing AI. The coolest part? It got better over time by playing against itself. Talk about a self-taught genius!
ML Predicts What You Want to Type
Ever wonder how your keyboard magically suggests the next word as you type? That’s ML learning from your typing habits! (And yes, it’s also why your phone sometimes autocorrects to something wildly off—it’s still learning, okay?)
ML Beat a Human Poker Champion
In 2017, an AI named Libratus outplayed some of the world’s best poker players. It didn’t just crunch numbers—it bluffed and strategized like a pro. Who knew machines could be so sneaky?
Comments