The Master of Science in Artificial Intelligence program is offered through the College of Engineering and Computer Science at the University of Michigan – Dearborn. The program is 30 credits with each course worth three credits. Let’s take a look at the structure.

Capstone Options

There are two ways to organize the program. The first is through a Master’s Thesis and the other is through a Master’s Project. These two options make your course selection slightly different.

Master’s Thesis Option

• Foundation courses – 12 credit hours
• Concentration courses – 9 credit hours
• Elective courses – 3 credit hours
• Master’s Thesis — 6 credit hours

Mater’s Project Option

• Foundation courses – 12 credit hours
• Concentration courses – 9 credit hours
• Elective courses — 6 credit hours
• Master’s project – 3 credit hours

Comparison

Three are two major differences that stand out. The first is the amount and type of work of the capstone and the second is the number of elective courses that are available. The Thesis seems to be better if you’re extremely focused on one aspect of AI and plan to go on and earn a Ph.D. The Project might be better for someone who is still exploring the field, wants to remain a generalist, or wants to simply work in the private sector. If you’re interested in a lot of the courses, as I am, then the Project gives you that option. As you might have guessed, I’m choosing the Project where I will build my Perception Engine.

Concentrations

Three are four concentrations to choose from and this is a requirement. While I wish they had a more general option, the concentrations offered work. The options are Computer Vision (CV), Intelligent Interaction (II), Machine Learning (ML), and Knowledge Management and Reasoning (KMR).

Computer Vision

This concentration puzzles me a bit. The courses offered include computer graphics, data visualization, and digital image processing. These three groups alone account for over half of the courses in the concentration. While interesting, there are only a handful of courses (basically the remaining ones) that feel like they should be part of an AI degree to me. That said, three are courses I’m interested in here, as they apply to my Project. I just wish they offered more with object recognition.

Intelligent Interaction

This concentration is really just a combination of game programming and robotics. There are a couple of user interface courses thrown in, but not enough to select the concentration. A couple of courses seem interesting to me, and I may select one of them, but this concentration seems lacking to me.

Machine Learning

I have a feeling that this will be the most popular concentration. While it has the second-fewest courses to choose from, they all seem to work together to teach machine learning rather than a combination of two or three quasi-related groups like the other concentrations. My problem is that there are too many courses here that I want to take; I feel I need to take this concentration to be able to get in as much as possible. As you might guess, this is my current concentration.

Knowledge Management and Reasoning

This is probably the most hodgepodge concentration. It contains the only dedicated natural language processing course, some text and data mining courses, one course about expert systems, and three math-focused courses. There is a lot to choose from here, but I’m guessing it’ll be the least chosen concentration. The NLP course will likely be popular, but they could have done more with it.

Personal Thoughts

I’m honestly surprised at the choices in concentrations. I would have expected Machine Learning, Computer Vision, Natural Language Processing, Symbolic AI, and then a list of possible electives to choose from (or courses from a different concentration). Maybe these concentrations would be exactly three courses that teach the core knowledge and each syllabus would have recommended follow-up reading.

Core Courses

There are four Core courses: two are computer science-specific (algorithms and software engineering), one is a general AI course, and the last has a choice between a software-specific or engineering-specific AI course. The software-specific Core course is also listed under the Machine Learning concentration, but my current understanding is that it does NOT double count.