Li Hongyi machine learning 2020 notes (1)

一、Course Introduction

1. Machine learning is to find functions automatically

 

Regression: output is a scalar.

Classification：yes or no

　　Binary classification: RNN assignment

　　Multi-class Classification: CNN operations  

Generation: Generate structured and complex things (text pictures, etc.)

　　Seq2Seq, GAN operations

 

2、Supervised Learning: Labeled Data

 

 

 

unsuperived learning 

 

 3. Frontier research

Explainable AI

 

 Adversarial Attack: fight against attacks

 

Network Compression

Anomaly Detection: Anomaly detection (be able to know if you do not know)

 

Transfer Learning（Domain Adversarial Learning）

 

 

 Meta Learning：learn to learn

 

 Life-long learning: Lifelong learning

 

 

二、Regression

1. Application

 

 

 

 2、Step