{
  "id": "1909.12598",
  "version": "v1",
  "published": "2019-09-27T10:23:42.000Z",
  "updated": "2019-09-27T10:23:42.000Z",
  "title": "\"Best-of-Many-Samples\" Distribution Matching",
  "authors": [
    "Apratim Bhattacharyya",
    "Mario Fritz",
    "Bernt Schiele"
  ],
  "categories": [
    "cs.LG",
    "stat.ML"
  ],
  "abstract": "Generative Adversarial Networks (GANs) can achieve state-of-the-art sample quality in generative modelling tasks but suffer from the mode collapse problem. Variational Autoencoders (VAE) on the other hand explicitly maximize a reconstruction-based data log-likelihood forcing it to cover all modes, but suffer from poorer sample quality. Recent works have proposed hybrid VAE-GAN frameworks which integrate a GAN-based synthetic likelihood to the VAE objective to address both the mode collapse and sample quality issues, with limited success. This is because the VAE objective forces a trade-off between the data log-likelihood and divergence to the latent prior. The synthetic likelihood ratio term also shows instability during training. We propose a novel objective with a \"Best-of-Many-Samples\" reconstruction cost and a stable direct estimate of the synthetic likelihood. This enables our hybrid VAE-GAN framework to achieve high data log-likelihood and low divergence to the latent prior at the same time and shows significant improvement over both hybrid VAE-GANS and plain GANs in mode coverage and quality.",
  "revisions": [
    {
      "version": "v1",
      "updated": "2019-09-27T10:23:42.000Z"
    }
  ],
  "analyses": {
    "keywords": [
      "distribution matching",
      "hybrid vae-gan framework",
      "best-of-many-samples",
      "achieve state-of-the-art sample quality",
      "mode collapse"
    ],
    "note": {
      "typesetting": "TeX",
      "pages": 0,
      "language": "en",
      "license": "arXiv",
      "status": "editable"
    }
  }
}