ML Paper Challenge Day 40 — Generative Adversarial Nets

Papers with Code - Generative Adversarial Networks

Day 40: 2020.05.23
Paper: Generative Adversarial Nets
Category: Model/Unsupervised Learning

Generative Adversarial Nets

• a new framework for estimating generative models via an adversarial process —a minimax two-player game
• simultaneously train two models:
• a generative model G that captures the data distribution
• a discriminative model D that estimates the probability that a sample came from the training data rather than G.
• training procedure for G is to maximize the probability of D making a mistake.
• the entire system can be trained with backpropagation

Advantage

• Markov chains are never needed, only backprop is used to obtain gradients, no inference is needed during learning, and a wide variety of functions can be…