{
  "id": "1906.06419",
  "version": "v1",
  "published": "2019-06-14T21:58:58.000Z",
  "updated": "2019-06-14T21:58:58.000Z",
  "title": "Learning Correlated Latent Representations with Adaptive Priors",
  "authors": [
    "Da Tang",
    "Dawen Liang",
    "Nicholas Ruozzi",
    "Tony Jebara"
  ],
  "comment": "12pages, 2 figures",
  "categories": [
    "cs.LG",
    "stat.ML"
  ],
  "abstract": "Variational Auto-Encoders (VAEs) have been widely applied for learning compact low-dimensional latent representations for high-dimensional data. When the correlation structure among data points is available, previous work proposed Correlated Variational Auto-Encoders (CVAEs) which employ a structured mixture model as prior and a structured variational posterior for each mixture component to enforce the learned latent representations to follow the same correlation structure. However, as we demonstrate in this paper, such a choice can not guarantee that CVAEs can capture all of the correlations. Furthermore, it prevents us from obtaining a tractable joint and marginal variational distribution. To address these issues, we propose Adaptive Correlated Variational Auto-Encoders (ACVAEs), which apply an adaptive prior distribution that can be adjusted during training, and learn a tractable joint distribution via a saddle-point optimization procedure. Its tractable form also enables further refinement with belief propagation. Experimental results on two real datasets show that ACVAEs outperform other benchmarks significantly.",
  "revisions": [
    {
      "version": "v1",
      "updated": "2019-06-14T21:58:58.000Z"
    }
  ],
  "analyses": {
    "keywords": [
      "learning correlated latent representations",
      "adaptive prior",
      "correlated variational auto-encoders",
      "correlation structure",
      "learning compact low-dimensional latent representations"
    ],
    "note": {
      "typesetting": "TeX",
      "pages": 12,
      "language": "en",
      "license": "arXiv",
      "status": "editable"
    }
  }
}