IID-GAN: an IID Sampling Perspective for Regularizing Mode Collapse
Not on the map. Why?
Abstract (arXiv)
Despite its success, generative adversarial networks (GANs) still suffer from mode collapse, i.e., the generator can only map latent variables to a partial set of modes in the target distribution. In this paper, we analyze and seek to regularize this issue with an independent and identically distributed (IID) sampling perspective and emphasize that holding the IID property referring to the target distribution for generation can naturally avoid mode collapse. This is based on the basic IID assumption for real data in machine learning. However, though the source samples {z} obey IID, the generations {G(z)} may not necessarily be IID sampling from the target distribution. Based on this observation, considering a necessary condition of IID generation that the inverse samples from target data should also be IID in the source distribution, we propose a new loss to encourage the closeness between inverse samples of real data and the Gaussian source in latent space to regularize the generation to be IID from the target distribution. Experiments on both synthetic and real-world data show the effectiveness of our model.
Key facts. Cited by 1 paper in the map (data through 7 Oct 2026). Builds on 33 of its 57 references that are in the map. First posted on arXiv 1 Jun 2021.
Builds on
33 of 57 references.
References · 33
- 2014Deep Learning Face Attributes in the Wildcited by 3,679
- 2017Progressive Growing of GANs for Improved Quality, Stability, and Variationcited by 2,967
- 2016Improved Techniques for Training GANscited by 2,938
- 2017Improved Training of Wasserstein GANscited by 2,223
- 2015Variational Inference with Normalizing Flowscited by 2,161
- 2014Conditional Generative Adversarial Netscited by 2,152
- 2015Unsupervised Representation Learning with Deep Convolutional Generative Adversarial Networkscited by 2,048
- 2018Spectral Normalization for Generative Adversarial Networkscited by 1,618
- 2015LSUN: Construction of a Large-scale Image Dataset using Deep Learning with Humans in the Loopcited by 1,350
- 2020Training Generative Adversarial Networks with Limited Datacited by 1,184
- 2018Challenging Common Assumptions in the Unsupervised Learning of Disentangled Representationscited by 929
- 2018Computational Optimal Transportcited by 765
21 more
- 2018Which Training Methods for GANs do actually Converge?cited by 672
- 2018Disentangling by Factorisingcited by 614
- 2016Adversarial Feature Learningcited by 350
- 2019Large Scale Adversarial Representation Learningcited by 238
- 2016Coupled Generative Adversarial Networkscited by 216
- 2016Unrolled Generative Adversarial Networkscited by 208
- 2017VEEGAN: Reducing Mode Collapse in GANs using Implicit Variational Learningcited by 199
- 2019Mode Seeking Generative Adversarial Networks for Diverse Image Synthesiscited by 138
- 2016Mode Regularized Generative Adversarial Networkscited by 123
- 2019Diversity-Sensitive Conditional Generative Adversarial Networkscited by 77
- 2018On GANs and GMMscited by 62
- 2016Generative Multi-Adversarial Networkscited by 60
- 2017Multi-Agent Diverse Generative Adversarial Networkscited by 59
- 2017Dual Discriminator Generative Adversarial Netscited by 55
- 2020Diverse Image Generation via Self-Conditioned GANscited by 54
- 2017Variational Approaches for Auto-Encoding Generative Adversarial Networkscited by 52
- 2018Dist-GAN: An Improved GAN using Distance Constraintscited by 26
- 2018GDPP: Learning Diverse Generations Using Determinantal Point Processcited by 26
- 2019Prescribed Generative Adversarial Networkscited by 23
- 2018MGGAN: Solving Mode Collapse using Manifold Guided Trainingcited by 22
- 2020The Bures Metric for Generative Adversarial Networkscited by 1
Built on by
Citing papers per year: 2024: 1.
1 paper.