The big idea
⚙️
Data in, patterns out
A model learns from data during training, then applies what it learned during inference.
Artificial Intelligence · Lesson 4
How Artificial Intelligence Works
beginner⏱ 18 min · 8 cards
The lifecycle
How a system comes to life
Data → training → model → inference.
Compare
Training vs. inference
Training
Learning
The model adjusts to fit examples.
tap for example ↻
Inference
Using
The trained model answers new inputs.
tap for example ↻
Key term
Key term
What is a “model”?
tap to reveal ↻
Model
The set of learned patterns (parameters) that maps inputs to outputs after training.
Nuance
A word from the Maester
Maester Note
Garbage in, garbage out: a model is only as good as the data it learned from.
Recap
Key takeaways
- ✓AI development is a managed, iterative lifecycle.
- ✓Problem definition is the step most often skipped and most costly to skip.
- ✓Data preparation is the bulk of the work and the source of most failures.
- ✓Training teaches a model; inference uses it.
- ✓Evaluation on unseen data is essential, and one accuracy number is not enough.
- ✓Deployed performance decays as conditions change.
Quick check
Which statement best describes artificial intelligence?
You did it
Nice work
Mark this lesson complete to track your progress.