Unsupervised Image Generation with Infinite Generative Adversarial Networks

Abstract

Image generation has been heavily investigated in computer vision, where one core research challenge is to generate images from arbitrarily complex distributions with little supervision. Generative Adversarial Networks (GANs) as an implicit approach have achieved great successes in this direction and therefore been employed widely. However, GANs are known to suffer from issues such as mode collapse, non-structured latent space, being unable to compute likelihoods, etc. In this paper, we propose a new unsupervised non-parametric method named mixture of infinite conditional GANs or MIC-GANs, to tackle several GAN issues together, aiming for image generation with parsimonious prior knowledge. Through comprehensive evaluations across different datasets, we show that MIC-GANs are effective in structuring the latent space and avoiding mode collapse, and outperform state-of-the-art methods. MIC-GANs are adaptive, versatile, and robust. They offer a promising solution to several well-known GAN issues. Code available: github.com/yinghdb/MICGANs.

Habilitando Anotações De Dados Autônomos: Uma Abordagem De Aprendizado Por Reforço Com Humano No Loop

“…Para reduzir os esforços na aquisição de novas anotações, pesquisadores exploram novas abordagens [116] através de aprendizagens semisupervisionadas (semi-supervised) [78], supervisão fraca (Weak-Supervised) [104], não supervisionado (Unsupervised) [119], auto-supervisionado (Self-Supervised) [95], aumento de dados (Data augmentation) [102] e aprendizado ativo (active learning), este último descrito na Seção 2.4.…”