EECS 649 Introduction to Artificial Intelligence
Exam
Electronic Blackboard Submission Due: April 24, 2019 @ 9PM
Paper Copy Due: April 25, 2019 @ 4PM
200 Points
Directions: You must read and follow these directions carefully. This is a 3 hour open notes,
open internet exam. You may not collaborate or communicate with another student in the class
or outside the class regarding the exam from 6pm, 4/24/2019 to 5 AM, 4/25/2019.
In order to discourage cheating, if you can prove another student was asking you for help or
discussing the exam with you prior to you or them submitting the exam and are the first to report
it, you will receive 50 extra points (not to exceed 200 points total) and they will have 50 points
subtracted from their total). The person that attempted to cheat first will not be able to gain any
points for reporting another cheating student but the reported cheating student will be deducted
50 points. Any group attempt to maximize total or individual gains from any type of cheating
will all receive zero points for their exam grade. Any detected cheating or plagiarism will result
in a zero points exam grade.
There will be 5% grade reduction if the exam is turned in electronically after midnight and an
additional 5% grade reduction if turned in electronically after 2 AM. Exams will not be given
ANY credit if submitted electronically after 5AM, April 25, 2019.
No questions will be answered by the instructor during the exam. If you run into any issues, do
your best to describe your assumptions or any discrepancies and solve the problem.
The exam answers that are not part of the programming portion should be submitted as a PDF
file. Clearly label the problem numbers, letters, and answers. Questions 1-4 should be typed
(you may include diagrams if you wish). Question 5 can be submitted as a scanned PDF of a
handwritten answer since it involves“drawing”some diagrams. The programming portion
(Question 4) should be submitted as a zip file containing all of the requested data and code.
Make sure that your name is included at the top of each submitted page or file.
TURN IN A PAPER COPY OF YOUR EXAM ANSWERS IDENTICAL TO YOUR
ELECTRONICALLY SUBMITTED ANSWERS (MINUS THE DATA FILES) TO THE EECS
OFFICE IN EATON 2001 BY 4PM ON APRIL 25, 2019. NO LATE PAPER COPIES WILL
BE ACCEPTED.
First Name: __ Last Name: _
2
- [60 points] General Artificial Intelligence: Write a coherent and well organized one to two
page essay in paragraph form that explains what AI is, what it is not, and what are its
limitations or dangers. Be sure to include, explain, and clearly identify (e.g. number) the
following concepts: the history of AI, the present status of AI, intelligent agents and their
various architectures, problem solving as search, learning, environment characteristics of an
intelligent agent, ethics, real-world examples, and list some of the subfields of AI. The real
world examples you provide should be from the in-class guest lectures. Be sure to give your
essay a title and adhere to spelling and grammar rules. - [20 points] Logistic regression and deep learning: Briefly compare and contrast logistic
regression and deep learning. Be sure to give definitions of each. Be sure to adhere to
spelling and grammar rules. - [20 points] Reinforcement learning: Briefly explain what reinforcement learning is and how
does it relate to other methods of learning. Be sure to adhere to spelling and grammar rules. - [80 points] Programming Machine Learning: Write a program in the language of your choice
(e.g. R) to create a supervised learning model to predict the housing prices given the data
provided on Blackboard (housetrain.csv, housetest.csv, and housedata_description.txt).
Prepare the training set and test set to include only the following features: year and month of
sale, lot square footage, and number of bedrooms.
housetrain.csv – the training set
housetest.csv – the test set
housedata_description.txt – full description of each column
a. What is the particular supervised learning method you are using and why did you
choose it over other methods?
b. What did you do with the data to prepare it for processing? [Rename your prepped
data to housetrain_prepped.csv and housetest_prepped.csv]
c. How did you go about training your machine learning model (i.e. explain the stepby-step
process you used)?
d. What are the set of features and the specific coefficient values that give you the
best results?
e. What is your R2 value (i.e. r-squared value) and what does it mean?
f. What is your RMSE value between the logarithm of the predicted value and the
logarithm of the observed sales price. (Taking logs means that errors in predicting
expensive houses and cheap houses will affect the result equally.)
First Name: __ Last Name: _
3
g. What is the predicted sales price values for id 1625 in the housetest.csv file ? To
find the feature, or predictor, values for this problem, open the housetest.csv file
and look at the row that has id 1625. Use only the year and month of sale, lot
square footage, and number of bedrooms in your learning model from that
example.
h. Graduate students only: Create another learning model and determine the set of
features that provides the best result. List the features you identify and the
coefficients . State the R2 , RMSE values and compare them with the first
learning model you created. Graduate student with the best performing model
wins“the prize.”Undergraduate students may do this portion for extra credit.
i. Graduate students only: Write down the predicted sales prices for id 1625 in
test.csv using your new learning model from part h . Undergraduate students
may do this portion for extra credit.
j. Be sure to upload your R code and data to blackboard in a zip file labeled
<your name>_machinelearningcode.zip
Your R code and data should produce and display the results you describe in“a”
through“i”above.