Class Announcements
TA hour: Monday 14:20~17:20 at Lab 310, CSIE building
Note1: TA hours starts in October. (There are no office hours in September.)
2018.09.19 updated:
Please send an email to TA (r06922009 [at] ntu . edu . tw) to confirm that you will take this course.
The title of the email should be "[NeuralNetworks]<your name> <student ID number> taking this course".
TA will send the class materials and references to you then.
確定要修課或是旁聽的同學，請以會固定收信的email信箱寄信給助教(r06922009 [at] ntu . edu . tw)
標題煩請以如下格式:
[NeuralNetworks]蔡侑儒 r06922009 確定修課
[NeuralNetworks]蔡侑儒 r06922009 確定旁聽
助教收到信後，會將課程資料寄給同學。
Course Outline
- Fundamental Concepts and Models of mental processes
- Single-Layer Perceptron
- Multilayer Perceptron
- Hopfield model
- Recurent Network
- Associative Memories
- Self-Organizing Networks
- Reinforcement learning
Course Slides
- Chapter 1
Reification of Boolean Logic (please re-download this lecture note if you got it before 2006/09/30)
(lecture note: pdf)(exercise: pdf) - Chapter 2
Solution Space and Learning Behavior of McCulloch-Pitts Neuron
(lecture note: pdf, corrections in Chapter2.pdf : Fig.3 (left figure), F0 --> F15 (four black circles) ; Fig.3 (right figure), F15 --> F0 (four empty circles))(refer: J.19, B.18, B.26) (demo) (matlab code) (remark)
- Chapter 3
Learning with Quadratic Sigmoid Function
(Supplementary material)
(Supplementary material 2 ppt) (uni-percetron)
(lecture note:pdf) (exercise: pdf) (refer: C.10) (matlab code) - Chapter 4
Hidden Tree in Multilayer Network
This tree shows the hidden details of the MLP. It can pinpoint the
local detailed errors in the MLP during BP training.
(lecture note: pdf) (exercise: pdf) (refer: C.12) (demo Hamming.zip) (demo AIR0510.zip) (remark) - Chapter 5
Internal Representations of Hidden Layers
There are five types to operate the SIR, Type I: SIR-SOM; (tutorial) Type II:
SIR-kernel; (tutorial and code) Type III: SIR-recurrent; Type IV: SIR-Hopfield; Type V:
SIR-module.
(lecture note:pdf) (exercise: pdf) (refer: C.20) (matlab code) (Codes for Figure 5, 6, 7) (Supplementary material ppt) - Chapter 6
Hairy model
(lecture note: pdf) (exercise: pdf) (refer: J.15) (matlab code) (tutorial+demo+code) - Chapter 7
Caianiello Neuronic Equations
(lecture note: pdf)(exercise: pdf) (refer: B.1) - Chapter 8
Caianiello Polygonal Inequality
(lecture note: pdf) - Appendix A (pdf)
- References (pdf)
Textbook
- Neural Networks: a comprehensive foundation, second edition, by Simon Haykin, Prentice Hall International, Inc., 1999