Introduction to AI Basics

Welcome to the world of Artificial Intelligence (AI)! You might have seen AI in movies or used it on your phone to unlock your screen with your face. But how does it actually work? AI isn't "magic"—it is built on three very important pillars: Data, Algorithms, and Computing Power. In these notes, we will explore these three parts and see how they work together to make computers "smart."

The Three Pillars of AI

Think of building an AI system like baking a professional cake. To get a great result, you need three things:

  • Data: These are your ingredients (flour, eggs, sugar).
  • Algorithms: This is your recipe (the instructions on what to do with the ingredients).
  • Computing Power: This is your kitchen equipment (the oven and the mixer that do the hard work).

If any of these three are missing or weak, the AI will not work well!

1. Data: The Fuel for AI

Data is the information that we give to an AI system so it can learn. AI systems process different types of data, including:

  • Text: Books, emails, or chat messages.
  • Images: Photos of cats, dogs, or traffic signs.
  • Audio: Recorded speech or music.

Why is Data important?

AI learns by looking at thousands or even millions of examples. For example, to teach an AI to recognize a "bicycle," you must show it thousands of pictures of bicycles.

The "Garbage In, Garbage Out" (GIGO) Principle:
This is a very important concept in Computer Literacy. If you give an AI "garbage" data (data that is wrong, messy, or biased), the AI will give you a "garbage" result. The correctness of data is essential for AI to be useful.

Quick Review: Without enough high-quality data, an AI cannot learn patterns accurately.

2. Algorithms: The Logic of AI

An Algorithm (演算法) is a set of predefined rules or step-by-step instructions used to solve a problem or complete a task. In AI, the algorithm tells the computer how to "learn" from the data.

What do AI algorithms do?

Instead of a human programmer writing a rule for every single possibility, AI algorithms allow the system to:

  • Find patterns: Like noticing that most "bicycles" have two wheels.
  • Make predictions: Deciding if a new photo is a bicycle or a car.
  • Optimize solutions: Finding the fastest way to solve a problem.

Did you know?
In your S2 and S3 programming lessons, you design simple algorithms using flow control and loops. AI algorithms are just much more complex versions of these logical steps!

3. Computing Power: The Engine of AI

Computing Power (算力) refers to the ability of computer hardware to perform calculations. AI requires a massive amount of "math" to be done very quickly.

Fundamental Hardware of AI

To provide this power, AI systems rely on specific hardware components:

  • CPU (Central Processing Unit): The "brain" of the computer that handles general tasks.
  • Specialized Processors: Because AI involves millions of tiny calculations at once, we often use special chips (like GPUs) that are much faster at doing math than a regular CPU.
  • Memory (RAM): This holds the data that the AI is currently "thinking" about so it can access it instantly.

Why do we need more power now?
In the past, we had algorithms and data, but our computers were too slow. Today, our hardware is powerful enough to process billions of operations per second, which is why AI has become so popular recently.

Analogy: A professional chef (Algorithm) can have the best recipe, but if they only have a tiny toaster oven (Low Computing Power), they can't bake a cake for \(1,000\) people!

Common Mistakes to Avoid

Mistake 1: Thinking AI is "smart" like a human.
Reality: AI is just a machine following Algorithms using Data and Computing Power. It doesn't "understand" things the way you do.

Mistake 2: Thinking you only need one of the three pillars.
Reality: You need all three. Even the best algorithm will fail if you don't have enough data or a fast enough computer to run it.

Key Takeaways Summary

  • AI performance depends on Data, Algorithms, and Computing Power.
  • Data is the information (text, images, audio) used for training.
  • Algorithms are the rules and logical steps the AI follows.
  • Computing Power is the hardware (like the CPU) that performs the calculations.
  • Garbage In, Garbage Out (GIGO) means that bad data leads to bad AI results.

Note: In the next chapter, we will look at "How AI systems process data and learn" to see exactly how these three pillars work together through Machine Learning!