Loading the lesson
Loading the lesson
Forward Deployed Engineer · Programming Foundations · Lesson 1 of 10
How a machine learns from examples, and what a language model does.
Read the short notes and pictures.Do everything inside a blue frame.
In ordinary software, a person writes the rules. In machine learning, a person supplies examples and the computer works out the rules.
Ordinary program
if "free money" in email: spamA person writes each rule by hand. It only knows what it was told.
Machine learning
10,000 emails marked spam or notThe computer finds the patterns that separate the two, including ones nobody thought to write.
You want a program that recognises handwritten digits. Which approach fits?
A model is built once, slowly, from data. After that it is used many times, quickly.
feed inproducesis used forThe chat assistants you have used are large language models. They were trained on a very large amount of text to do one thing: guess what comes next. Doing that well, one word after another, produces answers, summaries and code.
A language model states something false with total confidence. Why can that happen?
Model
The trained thing that takes an input and gives an output.
Training
Adjusting the model until it fits the examples.
Inference
Using the trained model on something new.
Prompt
The text you give a language model to tell it what you want.
Keep these