Introduction:
In a video called "The story from ML to Deep Learning to GenAI," creator Linda Vivah explains how computer brains have changed over time. She shows how we went from computers that needed humans to give them every single instruction to computers that can now create art and stories on their own. The video is a seed planted in your mind. But I have taken this video and created a longer explanation to help you appreciate the beauty of the concept she introduces.
.
.
Here is a simple look at the three big steps in this journey:
Step 1: Old-School Machine Learning (ML)
In the beginning, computers were good at finding patterns in "clean" data, like lists of numbers or organized spreadsheets. However, there was a problem: humans had to tell the computer exactly what to look for.
If you wanted a computer to recognize a stop sign, you had to program it to look for a red shape with eight sides. The computer wasn't "thinking"; it was just following a recipe. This worked for simple math, but it didn't work for the "messy" stuff on the internet, like random photos and videos.
Step 2: Deep Learning (DL)
Around 2012, things changed. Computers got much faster, and we had millions of labeled pictures to show them. This started the era of Deep Learning.
Instead of describing a stop sign to the computer, scientists just showed the computer a million different pictures of stop signs. The computer figured out what a stop sign looked like all by itself! For the next ten years, computers got really good at sorting photos and translating languages.
Step 3: Generative AI (GenAI)
The newest step is all about Creation. In 2017, a new way of building AI called a "Transformer" was invented.
Before this, AI was good at understanding things that already existed. Now, AI can actually make new things. It looks at a huge amount of data and predicts what should come next to make a new sentence or a new picture. This is how tools like ChatGPT work today.
How the Story Fits Together
Linda Vivah explains that these three steps are like "nesting dolls." Each new step lives inside the one before it. Machine learning led to deep learning, and deep learning led to the creative AI we use now. The story isn't over yet—the next chapter is being written right now.
Expanded Glossary: AI Words to Know
- Artificial Intelligence (AI): The broad idea of making computers act or think like humans.
- Machine Learning (ML): A way for computers to learn from data without being specifically programmed for every single task.
- Structured Data: Information that is organized and easy to read, like a list of names or a budget spreadsheet.
- Feature Engineering: When a human has to pick out the important "clues" (like colors or shapes) for a computer to follow.
- Deep Learning (DL): A type of machine learning that uses "neural networks" to learn from huge amounts of messy data.
- Neural Network: A computer system designed to work a bit like the human brain, using layers of "nodes" to process information.
- GPU (Graphics Processing Unit): A powerful computer chip that helps AI process lots of information very fast.
- ImageNet: A famous, giant collection of millions of pictures used to teach computers how to see.
- Generative AI (GenAI): AI that can create brand-new content, such as stories, poems, or digital paintings.
- Transformer: A special "blueprint" for AI that helps it understand the context and the relationship between words in a long sentence.
- Algorithm: A set of step-by-step rules or instructions a computer follows to solve a problem.