Deep Learning Basic

Introduction

Artificial Inteligence - Mimic human intelligence
Machine Learning - Data driven approach
Deep Learning - Neural Networks

Key components of Deep Learning

1. The data that the model can learn from
2. The model how to transform the data
3. The loss function that quantifies the badness of the model
4. The algorithm to adjust the parameters to minimize the loss

Data

Data depend on the type of the problem to solve

1. Classification
2. Semantic Segmentation
3. Detection
4. Pose Estimation
5. Visual QnA
   

Model

Loss

The loss function is a proxy of what we want to acheive
1. Regression Task - MSE
2. Classification Task - CE
3. Probabilistic Task - MLE

Optimization Algorithm

1. Dropout
2. Early stopping
3. k-fold validation
4. Weight decay
5. Batch normalization
6. MixUp
7. Ensemble
8. Bayeisan Optimization