/
Browse
/
Courses
/
Exploring the World of AI
/
Assignment Week 3
The World of AI
What Is AI?
Exercises
An Agent That Learns
Exercises
Behind the Scenes: Reinforcement Learning
Exercises
Traditional AI and Machine Learning
Exercises
Downloads
Quiz
More About How Machines Learn
Supervised Learning
Exercises
Behind the Scenes: Supervised Learning
Exercises
Unsupervised Learning
Exercises
Behind the Scenes: Unsupervised Learning
Exercises
Downloads
Quiz
The (Smaller and the) Bigger Picture
Can Machines Think?
Exercises
A Closer Look at Neural Networks
Exercises
AI and Society
Exercises
Additional Notes for Teachers
Downloads
Quiz
The World of AI
What Is AI?
Exercises
An Agent That Learns
Exercises
Behind the Scenes: Reinforcement Learning
Exercises
Traditional AI and Machine Learning
Exercises
Downloads
Quiz
More About How Machines Learn
Supervised Learning
Exercises
Behind the Scenes: Supervised Learning
Exercises
Unsupervised Learning
Exercises
Behind the Scenes: Unsupervised Learning
Exercises
Downloads
Quiz
The (Smaller and the) Bigger Picture
Can Machines Think?
Exercises
A Closer Look at Neural Networks
Exercises
AI and Society
Exercises
Additional Notes for Teachers
Downloads
Quiz
Knowledge quiz
It's time to put what you've learned to the test, get 8 right to pass this week.
1.
What does classification refer to in the context of supervised learning?
Choose the correct answer.
Assigning a label or category to data according to its characteristics
Identifying the key characteristics of certain classes
Extracting the characteristics of data
Extracting data from features
2.
A neural network has 5 inputs, two hidden layers and 4 outputs. If this network is used for classification, how many classes can you distinguish?
Choose the correct answer.
1
2
4
5
3.
Consider the following calculation for the output of an artificial neuron: 1 x 0.2 + 0 x 0.7 + (-2) x 2 + 5 x 0.5 + 5 = 3.7. What is the bias of the neuron?
Choose the correct answer.
5
0.5
-2
3.7
4.
What does training an artificial neural network mean in the case of supervised learning?
Choose the correct answer.
Automatically finding fitting weights and biases that model the relationship between input and output
Automatically finding the best activation function
Automatically determining the best number of neurons in each layer
Automatically determining the best labels to use for assignment
5.
What does the conversation with the chatbot Mitsuku show us?
There are two correct answers.
Mitsuku can take the previous chat history into account.
Mitsuku suggests she understands by asking counter-questions.
Mitsuku can answer even complex mathematical questions.
Mitsuku builds logic to respond to new problems as well.
6.
What is strong AI?
Choose the correct answer.
Software that covers the entire spectrum of human cognitive abilities
Software that solves problems faster than humans
Computers with artificial intelligence and very low power consumption
Software that is trained by machine learning
7.
What does the Turing test check?
Choose the correct answer.
Whether a computer demonstrates intelligent behavior
Whether a computer-generated voice sounds realistic
How many human characteristics a computer demonstrates
What the latency of a chat system with artificial intelligence is
8.
What does explainability mean in the context of an AI system?
Choose the correct answer.
For a particular decision, the key influencing factors are shown.
Criteria are shown for all decisions.
The system visualizes the structure of the corresponding neural network to the user.
Users interact with a chatbot that tells the customer that it is sorry.
9.
How does an adversarial attack on an AI system work?
Choose the correct answer.
The input is minimally modified until it is classified differently by the AI system.
The AI is retrained with differently labeled data.
Stop signs are laminated in such a way that cameras can no longer see the sign.
All labels are deleted from the training data.