Machine LearningWeekly Plans

Updated 

ML(Machie Learning)
4th Semester: Datamatician Course (Computer Science)
 

This is a preliminary schedule: be prepared for changes !
CHECK ALWAYS THE RESPECTIVE WEEK

 

Week

Subjects

Litterature
Recommendations/Links

Exercises/Solutions/Examples

5

Getting ready 
Zealand, Roskilde

 

 

Doonesbury 1-2
Superman
, Hitman...

 

Compulsory

(ML) Aurélion Gérion:

Hands-On Machine Learning

 

Background readings. An easy reader.

(ML Beginners) Oliver Theobald: ML Absolute Beginners

Buying Books  NOW
Winterbathing in the fjord

 

Link to E-Book (ML)

Link to Jupyter Notebook

 

Videos and online literature

(A rather conprehensuve list)

 

 

6

MICL

ML The Landscape

  • Types of ML
  • Challenges
  • Testing
  • Validation
  • Linear regression

Install necessary SW

  • Anaconda
  • Jupyter
  • Spyder3
  • SciKit

Python exercises


Literature/recommendations/Links

Slides/Videos

Chapter 1 Assignments
No. 1 - 14

 

Exercises

Ananconda Installation Guide

Python Basic No. 1

 

Jupyter Introduction
Panda Introduction


Importing DataSets

 

Homeworks

Homework1: Python Basic No. 2

 

Solutions

Python 1 Solution

Python 2 Solution

Housing Test

 

7

MICL

End-to-End ML Project

 

 

Python exercises continues

 

 

 

Literature/recommendations/Links

Slides/Videos

 

Exercises

Python Basic No. 2

Linear Regression
Regression Performance

Housing Ch. 2 A

 

 

Solutions

Linear Regression Solution

Regression Performance Solution

Linear Regression Smart Solution

 

Housing Ch. 2 A Solution

8
JEAN

End-toEnd ML Project

Literature/recommendations/Links

  • ML: Ch. 2, p 62-84
  • Useful ML links 2

Slides/Videos

Exercises
Linear Regression Standard

Linear Regression Missing Data

Housing No. 2 B Solution

Solutions
Linear Regression Standard Solution
Linear Regression Missing DataSolution

9
MICL

 

Classification

  • Types of Ckassification
  • Binary classification
  • Multi classification
  • Cross Validation
  • Confusion matrix
  • Specifity&Recall
  • ROC-AUC

 

 

 

Literature/recommendations/Links

Slides/Videos

Exercises from week 7-8
to be continued and finalized

Exercises
Classification Chapter 3 Questions

Classfication MNIST Exercise

Mandatory 1: Linear Regression
Mandatory 1: Data set

Solutions

 

 

 

 

10
MICL
JEAN

Mandatory assignment

 

Mandatory 1: Linear Regression
Mandatory 1: Data set
11
JEAN

School Corona closed
Thursday

   

12
MICL

JEAN

Online introduction
Follow up on mandatory

Lessons cancelled

   

13
JEAN

Online

 

Training Models

  • Different Training Models
  • Gradient Descent
  • Bacth and Mini-batch GD
  • Stochastic Gradient Descent
  • Ridge vs. Lasso
  • Early stopping


Literature/recommendations/Links

  • ML: Ch. 4, p. 111 - 142
  • Useful ML links 4

Slides/Videos

 

 

 

14
MICL
Online

Logistic Regression

  • Logistic concepts
  • Probability
  • Sigmoid the logistic function
  • Cost function
  • Multinomial Logistic Regression
  • Softmax
  • Iris example in SciKit

Literature/recommendations/Links

  • ML: Ch. 4, p. 142 - 152
  • ML Beginners: p. 52 - 58
  • Useful ML links 4

Slides/Videos


Exercises
Logistic Regression Questions Chapter 4

Logistic Regression Iris Exercise
Logistic Regression Iris Program

14
MICL
Online

Decision Trees

  • Decision trees structure
  • Gini impurity
  • Entropy
  • CART Cost function
  • Instability & sensitivity
  • Iris example in SciKit


Literature/recommendations/Links

  • ML: Ch. 6, p. 175 - 187
  • ML Beginners: p. 98 - 114
  • Useful ML links 6

Slides/Videos

Exercises

Decision Trees Questions Chapter 6

Decision Tree Iris Exercise
Decision Trees Iris Program

15

MICL
Extra Lessons

Tuesday 1-6
Wednesday 1-4

Online

Finalizing Decision trees

Support Vector Machines
(SVM)

  • SVM concepts
  • Linear SVM
  • Hard and soft margin
  • Non-linear SVM
  • Polynomial features
  • The kernel trick
  • Computational complexity
  • SVM regression
  • Iris example in SciKit

Literature/recommendations/Links

  • ML: Ch. 5, p. 153 - 164
  • (p. 165 - 173 extensive)
  • Useful ML links 5

Slides/Videos

Exercises

SVM Questions Chapter 5

SVM Iris Exercise
SVM Iris Program

15

Easter Vacation
Thursday - Monday

Holidays That means students catch up

  Teachers prepare !

16

JEAN
Online

Unsupervised Learning (UL)

  • UL Concepts
  • Clustering usage
  • K-Means
  • Anomaly detection
  • Image segmentation
  • Prepocessing
  • Semisupervised learning
  • (DBSCAN)
  • Customer  Case

Supervised teaching in Unsupervised Learning Techniques

Literature/recommendations/Links

  • ML: Ch. 9, p. 235 - 255
  • (p. 255 - 258 extensive)
  • Useful ML links 9

Slides/Videos

 

Exercises

Unsupervised Learning Customer Exercise
Customer Dataset on Kaggle

17
MICL
Online

Artificiak Neural Networks(ANN)

  • ANN principles
  • Perceptrons
  • Backpropagation
  • MutiLevelPerceptrons (MLP)
  • MLP Classification
  • MLP Regression
  • Tensorflow
  • Keras
  • 3 x cases
  • Playground.Tensorflow.org

 

Literature/recommendations/Links

Slides/Videos

Exercises

ANN Questions Chapter 10

Tensorflow Installation

Perceptron Iris Exercise
ANN Programs Ch. 10


MLP Classfication Fashion Exercise


Playground at Tensorflow Exercise

Problems problems

18
Synopsis writing

Synopsis tips and tricks
Here are some ideas:

Tensorflow applied
Tensorflow and Big Data
Tools investigations and Interpolation
Tools and Face Recognition
Security in ML
Ensemble learning
Random Forest
Dimensonal reduction
Tensorflow & low level Python
Convolutional Neural Networks
Generative Adversarial Networks
Visual attention
Online ML and IOT-data
Android with Neural Networks

https://www.kaggle.com/datasets
Voice regognition
Bird songs
Bird pictures
COVID19
+1000 more datasets !!!!

 

 

19

Synopsis writing Synopsis tips and tricks  

20

Synopsis writing Synopsis tips and tricks  

21

Synopsis writing Synopsis tips and tricks  

22

Synopsis writing / handin    

23

Synopsis hand in    

24

Rehearsal for exam

 

 

25

Oral Exam

Good luck !

 

Exam Roll List

Re-exam unknown

35

 5th Semester starts

 

 

 


 
 

Maintenance by micl@easj.dk