CSCI 335 - Artificial Intelligence

Fall 2017

Course Overview:
This course is an introduction to understanding and implementing computer systems that fall under the heading of "artificial intelligence". At the end of this course, you will be expected to be able to:

Develop software that exhibits some degree of intelligent behavior relative to a given domain
Understand and apply heuristic search
Determine whether a heuristic function is monotonic
Apply an adversarial search algorithm in the context of a computer game
Understand the basics of automated symbolic planning
Understand and apply the following machine learning algorithms
- Multi-layer perceptron (including autoencoders)
- Self-Organizing Map
- k-nearest-neighbor
- Naive Bayesian Classifier
- Decision tree learning (including random forests)
- Q-learning
Distinguish among supervised learning, reinforcement learning, and unsupervised learning
Employ scientific experiments to analyze the performance of an intelligent system
Assess the scientific basis for claims about artificial intelligence

Instructor: Dr. Gabriel Ferrer

Office Hours (M.C. Reynolds 312):
By appointment. To make an appointment with me, visit http://drferrer.youcanbook.me. From there, you can see my availability and select an appointment time.

Class Web Page: http://ozark.hendrix.edu/~ferrer/courses/335/

Lecture Time: B3 (1:15 - 2:30 pm, Tuesday/Thursday)

Final Exam Period: Thursday, December 7, 2:00-5:00 pm

Required Textbook: On Intelligence, by Jeff Hawkins and Sandra Blakeslee
This book is an overview of a neuroscientific theory of intelligence; it is not a computer science text. We will use its theory of intelligence as a basis for assessing the "intelligence" of the systems we develop.

Grading:
There are a total of 1,000 points available over the course of the semester. The thresholds for earning each letter grade are as follows:

Letter grade	Points to achieve
A	900
B	800
C	700
D	600

Here are the semester's assignments and the associated points for each:

Assignment	Total Value
Project 1	80
Project 2	80
Project 3	80
Project 4	80
Project 5	80
Project 6	80
Project 7	80
Project 8	80
Final Project	240
AI Essay	120

Programming Projects: You will develop your understanding of artificial intelligence by developing software that uses AI algorithms to achieve goals in a concrete domain. These assignments will use the Java programming language.

Each project (80 points total) will involve each of the following:

Programming assignment (20 points): Implement an AI algorithm for a given task.
Experimentation (20 points): Run a series of experiments using your implemented algorithm.
Progress report (5 points): Give a one-slide PowerPoint progress report during the class period prior to the project's due date. In this report, present the results of your experiments so far.
Paper (25): You will write a paper describing the results of your experiments.
Final presentation (10 points): Give a three-slide PowerPoint presentation describing the results from your paper.

Presentations: As mentioned above, for each programming project you will give two presentations: a progress report and a final report. These presentations will be graded for clarity of exposition, proper use of slides, and accuracy of content and analysis.

Revisions: After projects are returned, you are welcome to revise and resubmit your work. I will grade anew each submitted revision, and average the original and revised grades to produce a new grade for that assignment. Revisions may be submitted anytime until the start of the final exam period.

No late work will be accepted. Any work not submitted on time is a zero. However, you may submit a solution after the deadline to qualify under the revision policy. In effect, this means that late work can earn up to half credit.

AI Essay: An important component of the course is developing the ability to reason critically about claims that a system is "intelligent" and about computational theories of mind. To this end, you will compose an essay in which you analyze the relationship between the cognitive theory described in the textbook and the types of algorithms we have studied in the course.

Final Project: Towards the end of the semester, you will select a final project topic. You will develop an intelligent system that either extends a concept that we have explored this semester, or uses a concept that we have not covered. You will give an oral presentation of your project during the final exam period for this course. The project write-up will also be due at that time.

Disabilities: It is the policy of Hendrix College to accommodate students with disabilities, pursuant to federal and state law. Any student who needs accommodation in relation to a recognized disability should inform the instructor at the beginning of the course. Students should contact Julie Brown in Academic Support Services (505-2954; brownj@hendrix.edu) to begin the accommodation process.

Schedule: The anticipated schedule for the semester is below. The instructor reserves the right to alter the schedule as necessary during the semester.

Date	Day	Topic/Activity	Reading	Assignment Due
8/22	Tuesday	Introduction Search Algorithms	None	None
8/24	Thursday	Search Algorithms	On Int 9-22	None
8/29	Tuesday	Presentations	None	Project 1
8/31	Thursday	Planning	None	None
9/5	Tuesday	Planning	On Int 23-39	None
9/7	Thursday	Presentations	None	Project 2
9/12	Tuesday	Adversarial Search	None	None
9/14	Thursday	Adversarial Search	On Int 40-64	None
9/19	Tuesday	Presentations	None	Project 3
9/21	Thursday	Machine Learning: kNN and Naive Bayes	None	None
9/26	Tuesday	Machine Learning: kNN and Naive Bayes	On Int 65-84	None
9/28	Thursday	Presentations	None	Project 4
10/3	Tuesday	Neural Networks: Perceptrons	None	None
10/5	Thursday	Neural Networks: Perceptrons	On Int 85-105	None
10/10	Tuesday	Presentations	None	Project 5
10/12	Thursday	Fall Break No class	None	None
10/17	Tuesday	Neural Networks: Self-Organizing Map	None	None
10/19	Thursday	Neural Networks: Self-Organizing Map	On Int 106-125	None
10/24	Tuesday	Presentations	None	Project 6
10/26	Thursday	Q-Learning	None	None
10/31	Tuesday	Q-Learning	On Int 125-144	None
11/2	Thursday	Presentations	None	Project 7
11/7	Tuesday	Machine Learning: Decision Trees and Random Forests	None	Final Project Proposal
11/9	Thursday	Machine Learning: Decision Trees and Random Forests	On Int 144-164	None
11/14	Tuesday	Presentations	None	Project 8
11/16	Thursday	Recent advances in AI	On Int 164-176	Revised Final Project Proposal
11/21	Tuesday	No Class	None	None
11/23	Thursday	Thanksgiving No class	None	None
11/28	Tuesday	Progress Reports	On Int 177-205	Progress Report
11/30	Thursday	Retrospective	On Int 205-233	None
12/7	Thursday	Final Project Presentations	None	Final Project Report
12/13	Wednesday	None	None	Term Paper