Math Colloquia

1414

Date | Apr 08, 2021 |
---|---|

Speaker | 류경석 |

Dept. | 서울대학교 수리과학부 |

Room | 129-101 |

Time | 16:00-17:00 |

※ 줌(zoom) 병행

Zoom: https://snu-ac-kr.zoom.us/j/84678072601

회의 ID: 84678072601

Generative adversarial networks (GAN) are a widely used class of deep generative models, but their minimax training dynamics are not understood very well. In this work, we show that GANs with a 2-layer infinite-width generator and a 2-layer finite-width discriminator trained with stochastic gradient ascent-descent have no spurious stationary points. We then show that when the width of the generator is finite but wide, there are no spurious stationary points within a ball whose radius becomes arbitrarily large (to cover the entire parameter space) as the width goes to infinity.

TEL 02-880-5857,6530,6531 / FAX 02-887-4694