Recently, learning machine learning attracts attention. It is fun to try.
After acquiring a doctoral degree (physics) with statistical mechanics as the pillar, after studying it by changing its interest to the information science system, this is quite interesting.
Besides, data science including machine learning is compatible with the exact road of physics that faces experimental data.
I think that it will be good as a basic subject of undergraduate course.
You can not miss not only getting as a tool but also having fun to understand.
When a large number of degrees of freedom gather, it becomes an object of analysis with statistical mechanics as a clue,
You can bring a cut by physics into the information science methodology.
Perhaps quantum mechanics, perhaps that way, may develop further.
I'm planning a lecture that will inspire me to think about it.
To understand and to become familiar with machine leanring from the perspective of statistical mechanics.
machine learning, statistical mechanics
✔ Specialist skills | Intercultural skills | Communication skills | ✔ Critical thinking skills | ✔ Practical and/or problem-solving skills |
Mainly in the format of lectures.
Course schedule | Required learning | |
---|---|---|
Class 1 | Beyond least-squares | Beyond least-squares |
Class 2 | Approximation of functions and machine learing | Approximation of functions and machine learing |
Class 3 | Implementation of neural net using Python and Chainer | Implementation of neural net using Python and Chainer |
Class 4 | Deep learing and kernel method | Deep learing and kernel method |
Class 5 | Sparse modeling | Sparse modeling |
Class 6 | Statististical mechanics of information and spin glass theory | Statististical mechanics of information and spin glass theory |
Class 7 | Physics and machine learing | Physics and machine learing |
one
Distributed as appropriate.
Mainly by homework
Understanding of the basics of statistical mechanics.