Generative Models

• What I can not create, I do not understand

https://deepgenerativemodels.github.io/

Introduction

• What does it mean to learn a generative model
• generative model은 단순히 생성모델이 아니다

Suppose we have some images of dogs

We want to learn a probability distribution p(x) such that

• Generation : If we sample xnew ~ p(x), xnew should look like a dog

  • implicit models

• Density estimation :p(x) should be high if x look like a dog (어떤이미지의 확률을 계산함)

  이건 마치 image classification

  • explicit models

• Unsupervised representation learning

  특정 image가 어떤 특징을 가지고있는지를 학습