Reinforcement Learning for Finance

Chapter 5. Generated Data

In the proposed adversarial nets framework, the generative model is pitted against an adversary: a discriminative model that learns to determine whether a sample is from the model distribution or the data distribution. The generative model can be thought of as analogous to a team of counterfeiters, trying to produce fake currency and use it without detection, while the discriminative model is analogous to the police, trying to detect the counterfeit currency. Competition in this game drives both teams to improve their methods until the counterfeits are indistinguishable from the genuine articles.

Goodfellow et al. (2014)

In their seminal paper, Goodfellow et al. (2014) introduce generative adversarial nets (GANs) that rely on a so-called generator and discriminator. The generator is trained on a given data set. Its purpose is to generate data that is similar “in nature,” that is, in a statistical sense, to the original data set. The discriminator is trained to distinguish between samples from the original data set and samples generated by the generator. The goal is to train the generator to produce samples that the discriminator cannot distinguish from original samples.

Although this approach might sound relatively simple at first, it has seen a large number of breakthrough applications since its publication. There are GANs available nowadays that create images, paintings, cartoons, texts, poems, songs, computer code, and even videos that are hardly ...