Course Introduction
Place In the AI Curriculum
Before CS 3642, need:
- data structures and algorithms at undergraduate level.
- Above implies discrete math (sets, logic, combinatorics, proofs, graphs).
After CS 3642, ready for:
- CS 4267 Machine Learning
- CS 4277 Deep Learning
- AI application courses:
- CS 4732 Machine Vision
- CS 4742 Natural Language Processing
This course is a gateway to all of AI and machine learning.
Workload
According to KSU policy, a 3-hour course means
- 2.5 instructional hours per week,
- 5 hours of work outside of class.
So you should spend 75 hours outslide of class on this course. I use these guidelines when I design a course.
By the way, I've realized that professors at Georgia Tech DO NOT design their courses with these workoad guidelines in mind!
Work Breakdown
- 40-50 pages of reading per week, except exam weeks
- 4 Programming Assignments (PA)
- Semester project
- Midterm and Final Exams
Reading
- AIMA: Artificial Intelligence: A Modern Approach, 4ed, by Stuart Russell and Peter Norvig, https://aima.cs.berkeley.edu/. US or Global Edition.
- Best book on AI by a large margin.
- Used by all the top schools.
- Readings should take 2 hours per week for 13 weeks, 26 hours total.
- My slides are like lecture notes, i.e., self-contained summaries of the material.
- Instead of problem sets, I will provide study guides that you can use while you read.
Programming Assignments
- Should take 12 hours each, 48 hours total.
- When you build, you understand.
- Designed to prepare you to build AI systems.
If you add up all the time estimates for each work category, you get 74 hours. This is an average, of course.
Exams
- Each exam preceeded by an in-class review.
- Each exam designed to take 1 hour.
- Final exam covers the material since the midterm exam, i.e., is not comprehensive.
- Knowing the material in the study guides and lecture slides well should get you to the A level. Doing the reading as well should make an A comfortable.