Question 1 (1 point)

Saved

In AI, one primary advantage of deep learning models is they can memorize the data. Therefore, inference ability is not needed.

Question 1 options:

	True
	False

Question 2 (1 point)

Saved

0-1 logic is a traditional AI problem.

Question 2 options:

	True
	False

Question 3 (1 point)

Saved

Gradient descent is an optimization algorithm, and it can be used in linear classification.

Question 3 options:

	True
	False

Question 4 (1 point)

Saved

In machine learning, a large learning rate will result in faster convergence.

Question 4 options:

	True
	False

Question 5 (1 point)

Saved

The Iris dataset is a typical unstructured dataset.

Question 5 options:

	True
	False

Question 6 (1 point)

Saved

Supervised learning uses labeled data to train machine learning models.

Question 6 options:

	True
	False

Question 7 (1 point)

Saved

Logistic regression is a typical method for regression tasks.

Question 7 options:

	True
	False

Question 8 (1 point)

Saved

In soft-margin SVM, a small C value will result in over-fitting.

Question 8 options:

	True
	False

Question 9 (1 point)

Saved

Weight decay can be used as a regularizer to train machine learning models.

Question 9 options:

	True
	False

Question 10 (1 point)

Saved

The original hard-margin SVM is a typical method to solve non-linear classification problem.

Question 10 options:

	True
	False

Question 11 (1 point)

Saved

Which one is a typical machine learning problem?

Question 11 options:

	supervised learning
	unsupervised learning
	reinforcement learning
	all of them

Question 12 (1 point)

Saved

Assuming X stands for a data point with m features, which method can be used for feature scaling?

Question 12 options:

	X/m
	(X-) / , where and are mean and standard deviation of X
	X/ + , where and are mean and standard deviation of X
	2*X

Question 13 (1 point)

Saved

In logistic regression, which one can be used as the function ?

Question 13 options:

	e^(-z)
	1 / (1+e^(-z))
	1+e^(-z)
	1 / e^(-z)

Question 14 (1 point)

Saved

Which one CANNOT be used for linear classification problem?

Question 14 options:

	linear regression
	logistic regression
	linear support vector machine
	none of them

Question 15 (1 point)

Saved

In python sklearn package, which function is used to train a machine learning model, such as a SVM model?

Question 15 options:

	fit()
	search()
	train()
	get()

Question 16 (1 point)

Saved

In which area AI can be applied?

Question 16 options:

	Natural Language Processing
	Medical Image Analysis
	Customized Recommendation
	All of them

Question 17 (1 point)

Saved

The following areas are closely tied to AI except

Question 17 options:

	Deep Learning
	Data Science
	SQL database
	Machine Learning

Question 18 (1 point)

Saved

Which option is NOT a hyper-parameter?

Question 18 options:

	number of training epochs
	learning rate
	model weight
	All of these options are hyper-parameters

Question 19 (1 point)

Saved

In the hard-margin SVM, the optimization problem can be solved by

Question 19 options:

	Closed-form programming
	Step programming
	Quadratic programming
	Adaptive programming

Question 20 (1 point)

Saved

Which one is NOT an AI method?

Question 20 options:

	Web
	Neural network
	SVM
	Logistic regression