Essec\Faculty\Model\Contribution {#2216
  #_index: "academ_contributions"
  #_id: "14809"
  #_source: array:26 [
    "id" => "14809"
    "slug" => "diffusion-schrodinger-bridges-for-bayesian-computation"
    "yearMonth" => "2024-02"
    "year" => "2024"
    "title" => "Diffusion Schrödinger Bridges for Bayesian Computation"
    "description" => "HENG, J., DE BORTOLI, V. et DOUCET, A. (2024). Diffusion Schrödinger Bridges for Bayesian Computation. <i>Statistical Science: a Review Journal</i>, 39(1), pp. 90-99."
    "authors" => array:3 [
      0 => array:3 [
        "name" => "HENG Jeremy"
        "bid" => "B00760223"
        "slug" => "heng-jeremy"
      ]
      1 => array:1 [
        "name" => "DE BORTOLI Valentin"
      ]
      2 => array:1 [
        "name" => "DOUCET Arnaud"
      ]
    ]
    "ouvrage" => ""
    "keywords" => array:5 [
      0 => "Optimal transport"
      1 => "Schrödinger bridge"
      2 => "score matching"
      3 => "Stochastic differential equation"
      4 => "Time reversal"
    ]
    "updatedAt" => "2024-05-27 09:32:02"
    "publicationUrl" => "https://doi.org/10.1214/23-STS908"
    "publicationInfo" => array:3 [
      "pages" => "90-99"
      "volume" => "39"
      "number" => "1"
    ]
    "type" => array:2 [
      "fr" => "Articles"
      "en" => "Journal articles"
    ]
    "support_type" => array:2 [
      "fr" => "Revue scientifique"
      "en" => "Scientific journal"
    ]
    "countries" => array:2 [
      "fr" => null
      "en" => null
    ]
    "abstract" => array:2 [
      "fr" => "Denoising diffusion models are a novel class of generative models that have recently become extremely popular in machine learning. In this paper, we describe how such ideas can also be used to sample from posterior distributions and, more generally, any target distribution whose density is known up to a normalizing constant. The key idea is to consider a forward “noising” diffusion initialized at the target distribution, which “transports” this latter to a normal distribution for long diffusion times. The time reversal of this process, the “denoising” diffusion, thus “transports” the normal distribution to the target distribution and can be approximated so as to sample from the target. To accelerate simulation, we show how one can introduce and approximate a Schrödinger bridge between these two distributions, that is, a diffusion which transports the normal to the target in finite time."
      "en" => "Denoising diffusion models are a novel class of generative models that have recently become extremely popular in machine learning. In this paper, we describe how such ideas can also be used to sample from posterior distributions and, more generally, any target distribution whose density is known up to a normalizing constant. The key idea is to consider a forward “noising” diffusion initialized at the target distribution, which “transports” this latter to a normal distribution for long diffusion times. The time reversal of this process, the “denoising” diffusion, thus “transports” the normal distribution to the target distribution and can be approximated so as to sample from the target. To accelerate simulation, we show how one can introduce and approximate a Schrödinger bridge between these two distributions, that is, a diffusion which transports the normal to the target in finite time."
    ]
    "authors_fields" => array:2 [
      "fr" => "Systèmes d'Information, Data Analytics et Opérations"
      "en" => "Information Systems, Data Analytics and Operations"
    ]
    "indexedAt" => "2024-12-03T17:21:44.000Z"
    "docTitle" => "Diffusion Schrödinger Bridges for Bayesian Computation"
    "docSurtitle" => "Journal articles"
    "authorNames" => "<a href="/cv/heng-jeremy">HENG Jeremy</a>, DE BORTOLI Valentin, DOUCET Arnaud"
    "docDescription" => "<span class="document-property-authors">HENG Jeremy, DE BORTOLI Valentin, DOUCET Arnaud</span><br><span class="document-property-authors_fields">Information Systems, Data Analytics and Operations</span> | <span class="document-property-year">2024</span>"
    "keywordList" => "<a href="#">Optimal transport</a>, <a href="#">Schrödinger bridge</a>, <a href="#">score matching</a>, <a href="#">Stochastic differential equation</a>, <a href="#">Time reversal</a>"
    "docPreview" => "<b>Diffusion Schrödinger Bridges for Bayesian Computation</b><br><span>2024-02 | Journal articles </span>"
    "docType" => "research"
    "publicationLink" => "<a href="https://doi.org/10.1214/23-STS908" target="_blank">Diffusion Schrödinger Bridges for Bayesian Computation</a>"
  ]
  +lang: "en"
  +"_type": "_doc"
  +"_score": 8.697312
  +"parent": null
}