Method for Exploring Generative Adversarial Networks (GANs) via Automatically Generated Image Galleries

要旨

Generative Adversarial Networks (GANs) can automatically generate quality images from learned model parameters. However, it remains challenging to explore and objectively assess the quality of all possible images generated using a GAN. Currently, model creators evaluate their GANs via tedious visual examination of generated images sampled from narrow prior probability distributions on model parameters. Here, we introduce an interactive method to explore and sample quality images from GANs. Our first two user studies showed that participants can use the tool to explore a GAN and select quality images. Our third user study showed that images sampled from a posterior probability distribution using a Markov Chain Monte Carlo (MCMC) method on parameters of images collected in our first study resulted in on average higher quality and more diverse images than existing baselines. Our work enables principled qualitative GAN exploration and evaluation.

著者
Enhao Zhang
University of Michigan, Ann Arbor, Michigan, United States
Nikola Banovic
University of Michigan, Ann Arbor, Michigan, United States
DOI

10.1145/3411764.3445714

論文URL

https://doi.org/10.1145/3411764.3445714

動画

会議: CHI 2021

The ACM CHI Conference on Human Factors in Computing Systems (https://chi2021.acm.org/)

セッション: Computational AI Development and Explanation

[B] Paper Room 02, 2021-05-14 01:00:00~2021-05-14 03:00:00 / [C] Paper Room 02, 2021-05-14 09:00:00~2021-05-14 11:00:00 / [A] Paper Room 02, 2021-05-13 17:00:00~2021-05-13 19:00:00
Paper Room 02
12 件の発表
2021-05-14 01:00:00
2021-05-14 03:00:00
日本語まとめ
読み込み中…