NEW
Font size
WorksheetsHour of AI Videos from Code.org Review
Total questions: 18
Worksheet time: 9mins
What is the main reason the video says machine learning needs high-quality training data?
So the computer can run faster
So the computer can learn accurately
So people will watch more videos
So fewer images are needed
According to the video, why might a computer fail to recognize a disease in an X-ray?
Because X-rays are low resolution
Because the computer never learned how to read images
Because the training data may have come only from men
Because doctors refuse to label images
What does the video mean when it says "the data IS the code"?
That data can be written as computer code automatically
That selecting training data is like programming the algorithm
That data and code are stored in the same file
That data can replace human doctors entirely
According to the video, what does machine learning need to learn?
A list of steps written by hand
Lots and lots of data
A human to click every button
A special kind of battery
What did early computer scientists study to learn how to make machines that learn?
Plants
The human brain
Rocks
Cars
What is a neuron described as having in the video?
One middle part only
Three equal outputs
Two ends: inputs and a single output
No connections
In the movie recommendation example, how do the critics’ opinions start out?
Some critics count more than others right away
All critics’ opinions are ignored
All critics’ opinions carry the same weight at first
Only the newest critic counts
What happens after you give your own rating for a movie in the example?
The system deletes the critics’ ratings
The system changes the weights so it matches your preferences more
The system stops learning forever
The system adds a new critic automatically
Which of the following is NOT mentioned as a real-world use for neural networks?
Predicting floods and forest fires
Detecting and curing disease
Growing healthier food
Building physical robots by hand
What does "computer vision" help machines do?
Cook food
Interpret images and recognize objects
Play music
Write stories
According to the video, why can traditional programming fail when recognizing shapes like X and O?
Computers cannot read numbers
The definitions can be too strict and miss rotated or partial shapes
Shapes are always the same size
The computer runs out of memory
How does the computer improve its guesses during machine learning training?
By copying other computers
By getting shown the correct answer after each guess and learning from mistakes
By turning off and on again
By asking a human to draw the picture
Why might some face recognition systems have trouble seeing people of color, according to the video?
Because color photos are illegal
Because the systems were mostly trained with photos of white people
Because cameras can't capture faces at all
Because faces are always in black and white
What is one reason computer vision must be used carefully in real-world applications?
It always makes perfect decisions
It can depend on the quality and variety of its training data, which affects safety for tasks like detecting pedestrians or medical images
It never needs many examples to learn
It only works at night
What do chatbots like ChatGPT use to learn and make sentences?
A large language model trained on lots of text
a single picture
a list of famous quotes
a robot brain that thinks like a person
Why does the video say AI uses probabilities when choosing words?
Because it always guesses randomly with no information
Because it can see the future
beause probabilities help predict which word fits best from what it learned
because it repeats the same sentence every time
In the Shakespeare example what was a problem when the AI only looked at one letter at a time?
It could write perfect plays right away
It had no context, so the output was not helpful or correct
it learned to draw pictures instead of writing
it stopped working completely
Which of these is NOT one of the three important additions the video says modern systems use?
training on lots of internet text like Wikipedia and GitHub
looking at words instead of individual letters
a lot of human tuning to reduce bias and danger
using only one sentence from a play and ignoring everything else
