← Back to course
AI Explorers 6-8 / Week 02 / Tuesday
2/6
Week 02 · Sensors and Perception

Tuesday

From signals to meaning
// From raw signals to meaning
⏱ about 20 min

Tuesday: From Signals to Meaning

Raven prints the raw data from one telescope snapshot. It is a long page of numbers with no picture at all.

"This is what the sensor really sends," she says. "Just brightness values, one after another."

Rocket squints at the page. "There is a comet hiding in these numbers? I would never spot it."

Nova dims her lights and projects the same numbers as a grid of gray squares. A fuzzy shape with a tail appears.

"The signal did not change," she says. "What changed is that someone worked out what it means."

"That," Raven says, "is perception."

Perception

Perception is the process of getting meaning from sensory signals.

A signal says how much light hit each spot. Perception says what the light pattern probably shows.

For Sky Sorter, perception is the step that turns a grid of brightness values into a label.

Real AI does not perceive the way a person does. It finds meaning by working with data and patterns.

Light reaches the telescope
↓
The camera records a signal
↓
Software measures the signal
↓
Sky Sorter predicts a label
↓
A person checks the label
SIGNAL OR MEANING?
  • Read the question.
  • Tap your answer.
"The top-left spot received a lot of light." Is that a signal or a meaning?
"This card probably shows a comet." Is that a signal or a meaning?
"The microphone recorded a sound that rose and fell." Signal or meaning?
"Someone is saying hello." Signal or meaning?
FROM SKY TO LABEL
  • ?The camera records the light as a signal.
  • ?Software measures features of the signal.
  • ?Light from an object reaches the telescope.
  • ?Sky Sorter predicts a label from those features.
  • ?A person checks the label against the true answer.
WHY THIS EXERCISEPerception is a chain, and a mistake at any link can lead to a wrong label.

When perception goes wrong

Perception can fail even when the sensor works perfectly.

A faint comet might look like a dim star. A short satellite streak might look like a smudge.

Signals can also be noisy. Noise is unwanted signal, like a speck of dust on the lens.

This is one reason AI can be wrong, and why people keep checking its outputs.

StatementTrue or false?
Perception is getting meaning from sensory signals.?
If the sensor works perfectly, perception can never be wrong.?
A speck of dust on the lens could add noise to a signal.?
Real AI experiences the night sky the way a person does.?
WHY THIS EXERCISEKnowing where perception breaks helps you design better checks for Sky Sorter.
Getting meaning from signals is called ____.
Unwanted signal, like dust on a lens, is called ____.
In the flowchart, who makes the last check? Type one word.
WHY THIS EXERCISEA person checking outputs is how mistakes get caught.
Dust on a lens adds unwanted signal. What is unwanted signal called? Type one word.
WHY THIS EXERCISENoise can hide or fake a pattern, so it is one way perception goes wrong.
Try it
Close your eyes and listen for one minute.
Write each sound as a signal first (loud, soft, rising, steady), then add the meaning you perceived.

Well reasoned. Tomorrow's Explorer Lab makes you the sensor.

← Monday