Statistical Learning Theory, Part I

Sumio Watanabe


Lecture in Department of Mathematical and Computing Science, Tokyo Institute of Technology.

Prerequisite Knowledge : Linear algebra, defferentiation, integration, probability and statistcs are necessary.

For everyone who wants to take Statistical Learning Theory (MCS.T403)

Statistical learning theory is a subject that requires linear algebra, differentiation, integration, probability, and statistics as prerequisite knowledge. For this reason, an examination will be conducted to confirm basic knowledge in the first lecture (from 9:00 on June 13), and only successful applicants may apply for credits. Please note that those who are absent for the first time or who have forgotten writing materials can not take the exam.

(1) Basic concepts in Statistical Learning Theory
Lecture Note 1 , Classification by a neural network, mp4 file

(2) Neural Newtwork
Lecture Note 2 ,

(3) Neural Network
Lecture Note 3 ,

(4) Boltzmann Machine
Lecture Note 4 , Associative Memory, mp4 file .

(5) Deep Learning
Lecture Note 5 ,
Learning in Neural Network ,
Autoencoder, mp4 file ,
Convolution Neural Network, mp4 file ,

(6) Entropy and Information
Lecture Note 6 , Learning Curve, mp4 file ,

(7) Learning and Generalization
Lecture Note 7 , Generalization loss and information criteria in Neural Networks, mp4 file ,

(8) Knowledge Discovery
Lecture Note 8 ,
(Realizable Case) Model Selection by AIC and BIC, mp4 file ,
(Unrealizable Case) Model Selection by AIC and BIC, mp4 file ,