| Title |
Flow-based Generative Model using Invertible Transformer |
| Authors |
권세이(Se I Kwon) ; 최계원(Kae Won Choi) |
| DOI |
https://doi.org/10.5573/ieie.2023.60.12.79 |
| Keywords |
Generative model; Transformer; Normalizing flow; Conditional probability |
| Abstract |
Unlike Variational AutoEncoder(VAE), Generative Adversarial Network(GAN), a flow-based generative model explicitly learns the distribution of data x by a sequence of invertible transformation. This paper proposes a flow-based generative model using invertible transformer, which processes time-series data. We can estimate the distribution of real data from standard normal distribution using the proposed model. A learned model with condition generates new time-series data which follows the distribution of real data. |