Machine Learning

NTNU 機器學習

Back to Note Overview

Score

• Homework 30
• Quiz 5
• Q&A 10
• Examination 25
• Presentation 30

Note

Ch.01 Introduction

Traditional AI v.s. ML

• Traditional AI
  • 手動增加 explicit rules
• ML
  • 自動從大量範例中學習規則
• DL
  • using parallel simple algorithms to extract rules

Regularity

$\theta$: parameter

• $y=f(x)$
  • 連 form 不知道
• $y=f(x|\theta )$
  • 表示知道 form 不知道 parameter
  • 從 training data 裡面找到 parameter
• Linear Model
  • $y=a_{1}x+a_0,\theta =(a_0,a_1)$
• Quadratic model
  • $y=a_2{x}^2+a_{1}x+a_0,\theta =(a_0,a_1,a_2)$
• Gaussian model
  • $y=\frac{1}{\sqrt{2\pi }\sigma }\exp \frac{(x-\mu {)}^2}{2{\sigma }^2},\theta =(\sigma ,\mu {)}^T$

Ch.02 Supervised Learning

Ch.04 Parametric Methods

Ch.05 Multivariate Methods

Ch.06 Demensionality Reduction

Ch.09 Decision Trees

Ch.11 Multilayer Perceptrons

Ch.12 Deep Learning

Quiz

0914

• What kind of probability is often used to expressed associations among data?
• Regularities of data are often expressed as __.
• Address the difference between the expressions
  $f(x)$ and
  $f(x|q)$