The Basics, the review.
The Basics in twelve concepts and six new questions. If an answer doesn't add up, under the explanation you'll find the chapter to read again.
The key concepts.
- Artificial intelligence
The broad name for programs that do things that usually take a person, like recognizing an object, understanding a sentence or answering a question. Today almost all of it is built with machine learning.
- Machine learning
A way of building programs where nobody writes the rules by hand. You show many examples with the right answer and the program finds the rules on its own. Neural networks are the most common kind.
- Model
A program that learned from examples how to turn some data into an answer. Inside are the weights, the numbers it adjusted during training.
- Parameter
One of the numbers a model adjusts while it learns, also called a weight. MobileNet has about 3.5 million, a small language model from one to four billion.
- Inference
Using a model that's already trained to give an answer about new data. It's the work the phone does when it recognizes something.
- Classification
The task of picking one answer from a fixed list of classes by looking at the whole input. Healthy or sick, dog or cat, walking or running.
- Dataset
The set of examples you use to train and check a model. Photos, audio clips, sensor recordings or sentences, each one with its own label.
- Edge AI
Running a model on the device where the data is created, like a phone, a watch or a camera, instead of in a data center.
- Latency
The time between the question and the answer, measured in milliseconds. To keep up with a video at 30 frames per second it has to stay under 33 ms.
- NPU
Neural processing unit. The part of the phone's chip made specially for the multiplications in neural networks. It uses much less power than the CPU for the same work.
- RAM
The phone's working memory, from 6 to 16 GB. It holds the model's numbers while it runs. How fast they can be read from there often matters more than the power of the chip.
- Quantization
Writing the weights with simpler numbers, for example 8-bit integers instead of 32-bit decimals. The model gets about four times smaller and faster, and it usually loses little.
6 questions that weren't in the lessons.
They bring the course's lessons together. After each answer there's the explanation and the place to read again. The review doesn't change the map.
-
An app recognizes plants even high in the mountains, where the phone has no signal. What does that tell you?
If the answer arrives without a network, the phone does the math. That's Edge AI, and it's why it works in airplane mode too. Explained in 01.01, chapter 5 →
-
A model has 3 billion parameters, each one written with 2 bytes. The phone has 4 GB of RAM. Does it fit?
3 billion weights times 2 bytes make 6 GB, more than all the memory. First it has to be slimmed down, for example with quantization. Explained in 01.03, Go deeper →
-
During training, what changes after each step over the examples?
Learning means adjusting the weights. At each step the loss goes down a little, and after many epochs the model answers better. Explained in 01.03, chapter 6 →
-
On a video call, the phone blurs the background behind you and keeps your hair sharp. Which task is the model doing?
To blur only the background, it needs to know which pixels are yours, right up to the edge. A box would take in a piece of background too, a class would only say you're there. Explained in 01.04, chapter 4 →
-
You have 500 photos of leaves, all taken on one sunny afternoon. What's the risk?
The model only learns from what you show it. If the photos all look alike, it works in those conditions and badly in the others. Explained in 01.05, chapter 2 →
-
What sets machine learning apart from a program written by hand?
In machine learning nobody writes the rules. The program works them out from lots of examples with the right answer, and neural networks are just one way to do it. Explained in 01.04, Go deeper →