I have been teaching in Japanese university for eight years as assistant and associate professors.

Possible Cources for Undergraduate

1. information theory (introduction to data compression, error correcting codes, and cryptography)
2. artificial intelligence (introduction to machine learning and inference)
3. computation theory (NP-complete problems, mathematical foundation)
4. applied statistics (Bayesian and mathematical statistics)

Possible Cources for Graduate

1. data compression (universal data compression, MDL, connection with machine learning)
2. machine learning (Bayesian and neural networks, data mining)
3. evolutionary computation (theory, applications, programming with C++)
4. cryptography (public key cryptography basics, number theory problems with application to cryptography, algebraic curve cryptography, programming with Lidea, Magma, NTL, etc.)

Policy on Classes

1. Concise and full of examples
2. Intimate and welcome atmosphere.
3. At least one homework set for everyweek. Through the problems, the students should be able to check how well they understand the class. I ask the graders to hold problem sessions if the students request them.
4. Two examinations for one term.
5. Informative homepage presentation and mailing list for announcements and questions/answers.

For Graduate Students in Laboratory

For graduate students, I have many topics on information theory and artificial intelligence which I would accomplish if I had enough time. (If the students are willing to choose what they wish to do, I will not impose my preference on them.) Laboratory students are required to read books on the basic material in particular for the first year. Next, I suggest several topics which they seem to be interested in and discuss the outline of what they do as their Ph.D works. However, they must find their own problem. Otherwise, their thesis would be no more than execise problems. What I require most for students is to be independent from their supervisor before they get Ph.D.