gonken-lesson

• This repo contains the lesson in GonKen

Overviews

---

1. Introduction
2. Learning Algorithm
3. How to Evaluate Your Model
4. Overfitting and Underfitting
5. Practice: Overfitting and Underfitting
6. Bias and Variance / Hyperparameter and Validation Set
7. Maximum Likelihood Estimation
8. Practice: Unsupervised and Supervised learning
9. Introduction of Deeplearning
10. Gradient-Based Learning and SGD
11. Backpropagation
12. Practice: Neural Network Instruction with PyTorch
13. Regularization 1
14. Regularization 2
15. Practice: Overfitting and Underfitting
16. Practice: Regularization

Lessons

---

Introduction

• What are AI, machine learning, deeplearning.
• Why we should use deeplearning in certain tasks

Learning Algorithm

• 5.1
• discrete example for machine learning algorithm

How to Evaluate Your Model

• 11.1

Overfitting and Underfitting

• 5.2

Practice: Overfitting and Underfitting

• Linear/Poly Regression model
• Decision Tree

Bias and Variance / Hyperparameter and Validation Set

• 5.3
• 5.4.4

Maximum Likelihood Estimation

• math excercise

Practice: Unsupervised and Supervised learning

• unsupervised: PCA, Clastering(k-means), KDE
• supervised: random forest, logistic regression, kNN, NN

Introduction of Deeplearning

• 6

Gradient-Based Learning and SGD

• 4.3(no 4.3.1), 5.9

Backpropagation

• 6.3.1
• Why ReLU is employed as activation function in DNN

Practice: Neural Network Instruction with PyTorch

• Build simple structure of NN in PyTorch

Regularization 1

• 7.1, 7.4, 7.5, 7.8
• L1 and L2

Regularization 2

• 8.1.3, 8.7.1

Practice: Overfitting and Underfitting

• train in dataset which is easy to be overfitting/underfitting

Practice: Regularization

• to prevent overfitting