Essec\Faculty\Model\Contribution {#2216
  #_index: "academ_contributions"
  #_id: "14100"
  #_source: array:26 [
    "id" => "14100"
    "slug" => "computational-doob-h-transforms-for-online-filtering-of-discretely-observed-diffusions"
    "yearMonth" => "2023-07"
    "year" => "2023"
    "title" => "Computational Doob h-transforms for Online Filtering of Discretely Observed Diffusions"
    "description" => "CHOPIN, N., FULOP, A., HENG, J. et THIERY, A.H. (2023). Computational Doob h-transforms for Online Filtering of Discretely Observed Diffusions. Dans: <i>Proceedings of the 40th International Conference on Machine Learning, PMLR 202:5904-5923</i>. Honolulu:  Proceedings of Machine Learning Research."
    "authors" => array:4 [
      0 => array:3 [
        "name" => "FULOP Andras"
        "bid" => "B00072302"
        "slug" => "fulop-andras"
      ]
      1 => array:3 [
        "name" => "HENG Jeremy"
        "bid" => "B00760223"
        "slug" => "heng-jeremy"
      ]
      2 => array:1 [
        "name" => "CHOPIN Nicolas"
      ]
      3 => array:1 [
        "name" => "THIERY Alexandre H."
      ]
    ]
    "ouvrage" => "Proceedings of the 40th International Conference on Machine Learning, PMLR 202:5904-5923"
    "keywords" => array:10 [
      0 => "Computational Doob h-transforms"
      1 => "Online filtering"
      2 => "Discretely observed diffusions"
      3 => "Machine learning"
      4 => "Stochastic processes"
      5 => "Bayesian filtering"
      6 => "State estimation"
      7 => "Hidden Markov models"
      8 => "Sequential Monte Carlo methods"
      9 => "Probabilistic inference"
    ]
    "updatedAt" => "2024-10-31 13:51:19"
    "publicationUrl" => "https://proceedings.mlr.press/v202/"
    "publicationInfo" => array:3 [
      "pages" => null
      "volume" => "202"
      "number" => null
    ]
    "type" => array:2 [
      "fr" => "Actes d'une conférence"
      "en" => "Conference Proceedings"
    ]
    "support_type" => array:2 [
      "fr" => "Editeur"
      "en" => "Publisher"
    ]
    "countries" => array:2 [
      "fr" => "Royaume-Uni"
      "en" => "United Kingdom"
    ]
    "abstract" => array:2 [
      "fr" => "This paper is concerned with online filtering of discretely observed nonlinear diffusion processes. Our approach is based on the fully adapted auxiliary particle filter, which involves Doob’s htransforms that are typically intractable. We propose a computational framework to approximate these h-transforms by solving the underlying backward Kolmogorov equations using nonlinear Feynman-Kac formulas and neural networks. The methodology allows one to train a locally optimal particle filter prior to the data-assimilation procedure. Numerical experiments illustrate that the proposed approach can be orders of magnitude more efficient than state-of-the-art particle f ilters in the regime of highly informative observations, when the observations are extreme under the model, or if the state dimension is large."
      "en" => "This paper is concerned with online filtering of discretely observed nonlinear diffusion processes. Our approach is based on the fully adapted auxiliary particle filter, which involves Doob’s htransforms that are typically intractable. We propose a computational framework to approximate these h-transforms by solving the underlying backward Kolmogorov equations using nonlinear Feynman-Kac formulas and neural networks. The methodology allows one to train a locally optimal particle filter prior to the data-assimilation procedure. Numerical experiments illustrate that the proposed approach can be orders of magnitude more efficient than state-of-the-art particle f ilters in the regime of highly informative observations, when the observations are extreme under the model, or if the state dimension is large."
    ]
    "authors_fields" => array:2 [
      "fr" => "Systèmes d'Information, Data Analytics et Opérations"
      "en" => "Information Systems, Data Analytics and Operations"
    ]
    "indexedAt" => "2024-11-21T10:21:50.000Z"
    "docTitle" => "Computational Doob h-transforms for Online Filtering of Discretely Observed Diffusions"
    "docSurtitle" => "Conference Proceedings"
    "authorNames" => "<a href="/cv/fulop-andras">FULOP Andras</a>, <a href="/cv/heng-jeremy">HENG Jeremy</a>, CHOPIN Nicolas, THIERY Alexandre H."
    "docDescription" => "<span class="document-property-authors">FULOP Andras, HENG Jeremy, CHOPIN Nicolas, THIERY Alexandre H.</span><br><span class="document-property-authors_fields">Information Systems, Data Analytics and Operations</span> | <span class="document-property-year">2023</span>"
    "keywordList" => "<a href="#">Computational Doob h-transforms</a>, <a href="#">Online filtering</a>, <a href="#">Discretely observed diffusions</a>, <a href="#">Machine learning</a>, <a href="#">Stochastic processes</a>, <a href="#">Bayesian filtering</a>, <a href="#">State estimation</a>, <a href="#">Hidden Markov models</a>, <a href="#">Sequential Monte Carlo methods</a>, <a href="#">Probabilistic inference</a>"
    "docPreview" => "<b>Computational Doob h-transforms for Online Filtering of Discretely Observed Diffusions</b><br><span>2023-07 | Conference Proceedings </span>"
    "docType" => "research"
    "publicationLink" => "<a href="https://proceedings.mlr.press/v202/" target="_blank">Computational Doob h-transforms for Online Filtering of Discretely Observed Diffusions</a>"
  ]
  +lang: "en"
  +"_type": "_doc"
  +"_score": 8.21763
  +"parent": null
}