AI & Machine LearningPattern Recognition

Drawing the Line

Classifiers draw invisible boundaries between kinds of things

Your email app quietly decides, every single day, which messages are spam. It can't read the way you do. So how does it choose?

You'll find out

  • How an email turns into a dot
  • Your line against the machine's
  • Why pictures once fooled spam filters

Why it matters

Spam filters, face unlock and fraud alerts all sort things by drawing lines. Knowing how shows where they slip.

About 3 min6 interactive slides · 5 questionsLesson 305 · Zone 34

Play this lesson

Free to play in your browser or on iPhone

More in this zone

  1. 298Computational GeometryThe Nearest-Café MapDivide a city by which café is closest and you get Voronoi cells
  2. 299AerodynamicsWhy Golf Balls Have DimplesDimples help a golf ball fly much farther
  3. 300StatisticsHow a Poll of 1,000 Speaks for MillionsA good spoonful tells you about the whole pot
  4. 301StatisticsThe Poll With 2 Million Answers That Got It WrongIn 1936 a huge survey predicted the wrong US president
  5. 302FeedbackThe Microphone SquealSound chasing itself in a loop screams
  6. 303Web SecurityLittle Bobby TablesA name typed into a form can delete a whole database
  7. 304Communication HistoryWhat Hath God WroughtThe first long-distance telegraph message, in 1844
  8. 306Visual IllusionsThe Checker Shadow IllusionTwo squares are exactly the same gray