Many new students have attended this year the university. They are trying to learn something about artificial intelligence and have selected one of the leading universities in Europe and the United States. The individual biographies of the students are different. Most of them are male, but some female students are motivated to learn new things about computer science as well. From the perspective of a teacher the situation is problematic, because the students in the year 2019 are different from their counterparts 40 years ago. It has to do with the knowledge the students bring with them. In the past, it was rare, that a student was a proud owner of his own computer, and the internet wasn't invented yet. Today, that is the normal case. This effects the needs of students which topic they would like to learn.

What is the demand of a modern student in relationship to an Artificial Intelligence course? Has he a special subject which he is more interested than other? Does t`he average student asks for help in mathematics or is he interested in learning AI from a historical perspective?

Response to comment

According to the comment, the question at all doesn't make much sense, because it can't be answered by a hard formula from computer science or mathematics. I've taken this concern seriously and asked an academic search engine if “teaching AI at the university” is discussed in the literature or not. I've found many examples in which the problem was addressed in the current literature:

  • Kumar, Deepak, and Lisa Meeden. "A robot laboratory for teaching artificial intelligence." ACM SIGCSE Bulletin 30.1 (1998): 341-344.

In the paper, the question was not “what is AI?”, but the problem was how to teach the subject to students. In other papers, the Robocode environment was mentioned as a framework for teaching AI programming at the university. The conclusion is, that my original question is very ontopic in an AI forum and the criticism is subjective.

  • 1
    $\begingroup$ Several questions and subjective... This is not the forum for that I believe. Maybe try reddit? $\endgroup$ – Miguel Saraiva Sep 6 '19 at 19:17

As student and TA, I found that:

  1. Mathematical foundations, or at very least pointers on where to get them. Some students might find this not useful, but structuring the functioning of models in terms of mathematical constructs will help the student in later phases.

  2. Computational resources as Google colab, Kaggle kernels and the likes. Some people don't know they exist, but some others think of these as troublesome tools to get to work. Live demos always help with both.

  3. I don't know what are the subjects of greatest interests for students in general, but showing applications in the real world is of primary importance. Self-driving cars, applications of Reinforcement learning to robotics, object recognition applied to medical diagnosis, are all fields appealing to different kind of students with different vocations and ambitions. It would make sense to keep it varied not to preclude the engagement of any specific group.

  4. As you bring up the history, I find it as an essential part of my study and helpful to understand how the field evolved over time. This is of course personal, although I must say that all the lecturers in my AI master briefly referred to the history behind the specific method/model sooner or later.

Hope that helps.

  • 1
    $\begingroup$ Just my (possibly worthless) 2 cents; regarding point 3, I find that super annoying and boring, I'd rather listen to theoretical topics/challenges and the like. And of course I am not saying it is not a good idea to show applications (indeed, most people would find that interesting) but rather, don't forget to point at the deeper issues sometimes. I am posting this comment because often professors tend to overemphasize applications! $\endgroup$ – olinarr Sep 6 '19 at 15:38

There is no single answer applicable to all sorts of students, but definitely practical aspects would interest most of them. Aspects, that young students see modern companies or start-ups are trying to use to create succesfull product or service. As being a data science student myself I would say that historical perspective should be rather used only as an introduction to some serious subject.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.