I have heard in several media that the next great revolution will be that of Artificial Intelligence. And I agree. Nowdays, the problem of building software efficiently is in somehow solved. Today's universities should be anticipating to that inevitable, but necessary, fact. Specializations at the postgraduate level are not enough. There are AI issues that need to be addressed in more detail.
This topic is very close to many disciplines including:
So I strongly recommend to create and establish such important specialty, in return to your impressive question, such an engineer must be initiative, motivative, having excellent basis in mathematics, physics, and interested in computer programing, ...etc.
Regards, Emad
Actually, a lot of software is very inefficient, and only works because of the power of the computers running it. AI needs multidisciplinary teams, at the very least to ask the right questions, including ethical questions.
This topic is very close to many disciplines including:
So I strongly recommend to create and establish such important specialty, in return to your impressive question, such an engineer must be initiative, motivative, having excellent basis in mathematics, physics, and interested in computer programing, ...etc.
Regards, Emad
Degree totally depends on what do you expect or wait for your research. About skills, firstly candidates should have good math, statistics and electronic eng knowledge together with system programming. Secondly it requires good collaborations with interested fields to develope algorithms and systems.
AI has more to with math (probability /stats/L1/L2/Bayes network /MAP/Discrete Math /Linear algebra /Logic (may be from the age of Aristotle ) etc etc)
No hardcore confusion of mankind (Philosophy and Psychology) ....
The Question is a complex one. I am of the opinion that AI is more of a multidisciplinary application area than looking at a way of fusing or integrating it to a particular discipline and secondly AI is evolving which does not give it room to be study as a specialaized programme outside some core discipline where it is commonly used. AI is commonly used in the following discipline Computer Science, Electrical and Electronics Engineering, Industrial/System Engineering, Mechanical Engineering, Computer Engineering and Information Technology. Concerning skills needed or required for AI, they are good mathematics background, data structure background, statistics and good programming skills.
Hi Pavel,
What a great question!
It seems to me (as a multi-disciplinary engineer) that the parts of my various degrees that are most relevant to the usage of modern (deep learning, statistical ML, Numenta-like, RDF knowledge representation) AI were:
- Definitely more statistics than calculus, but calculus is still important for other engineering skills, so the degree should probably be rather heavy on math. Maybe 2 semesters of statistics and two of calculus?
- Discrete mathematics for the logic and introduction to networks.
- Algorithms and at least some programming courses from Computer Science
- The courses that make you an engineer. Those would be the ones common to electrical and mechanical engineering such as statics, dynamics, circuits. You could add to that (e.g. machine design or mechatronics) to fill out any extra room you have. You won't have much :)
Additionally, I think you would need most of these:
- Some kind of data organization class, e.g. from a data science program. So much of modern AI is arranging the data before input.
- Neural networks, basic and then deep learning.
- Philosophy of knowledge representation (ontology, particularly).
- A series of project courses that had the students build progressively more interesting models of real-world data, such as a basic neural network for recognizing letters, then a deep learning network to play games, perhaps a clustering project such as de-fuzzing license plate photos. Then you would need to define a capstone project, but that could be student-proposed and instructor-approved.
Please let me know if you continue with this project. It sounds fun!
Regards,
Dave
We cannot make a clear unique definition of AI. See good known book "Artificial Intelligence: A Modern Approach" of Russel and Norvig ". They give about 20 definitions. Now AI mostly becomes the problem knowledge domain, not subject one, and it ihas not hard frontiers.
We have to make difference between computational aspects and less formalized and opened component of AI, such as knowledge based systems, knowledge engineering etc., and, of course, we should keep in mind the existence of fundamental problem of understanding.
AI has mathematical, technical and humanitarian scientific basement and cross-disciplinary nature. So mixed specialisation would not be formalized very good from the view of educational standards, IMHO.
- Badges
- Science topic
- Similar topics
- Computer Science and Engineering
- Software