Conference
Προβολή σε EDS

Appa: Bending Weather Dynamics with Latent Diffusion Models for Global Data Assimilation

Λεπτομέρειες βιβλιογραφικής εγγραφής
Τίτλος: Appa: Bending Weather Dynamics with Latent Diffusion Models for Global Data Assimilation
Συγγραφείς: Andry, Gérôme, Lewin, Sacha, Rozet, François, Rochman, Omer, Mangeleer, Victor, Pirlet, Matthias, Faulx, Elise, Grégoire, Marilaure, Louppe, Gilles
Πηγή: Machine Learning and the Physical Sciences Workshop (NeurIPS 2025), San Diego, United States - California [US-CA], 06/12/2025
Έτος έκδοσης: 2025
Θεματικοί όροι: Computer Science - Learning, Physics - Atmospheric and Oceanic Physics, Engineering, computing & technology, Computer science, Ingénierie, informatique & technologie, Sciences informatiques
Περιγραφή: Deep learning has advanced weather forecasting, but accurate predictions first require identifying the current state of the atmosphere from observational data. In this work, we introduce Appa, a score-based data assimilation model generating global atmospheric trajectories at 0.25\si{\degree} resolution and 1-hour intervals. Powered by a 565M-parameter latent diffusion model trained on ERA5, Appa can be conditioned on arbitrary observations to infer plausible trajectories, without retraining. Our probabilistic framework handles reanalysis, filtering, and forecasting, within a single model, producing physically consistent reconstructions from various inputs. Results establish latent score-based data assimilation as a promising foundation for future global atmospheric modeling systems.
Τύπος εγγράφου: conference paper not in proceedings
http://purl.org/coar/resource_type/c_18cp
conferencePaper
peer reviewed
Γλώσσα: English
DOI: 10.48550/arXiv.2504.18720
Σύνδεσμος πρόσβασης: https://orbi.uliege.be/handle/2268/342181
Rights: open access
http://purl.org/coar/access_right/c_abf2
info:eu-repo/semantics/openAccess
Αριθμός Καταχώρησης: edsorb.342181
Βάση Δεδομένων: ORBi
FullText Text:
  Availability: 0
CustomLinks:
  – Url: https://orbi.uliege.be/handle/2268/342181#
    Name: EDS - ORBi (ns324271)
    Category: fullText
    Text: View record at ORBi
Header DbId: edsorb
DbLabel: ORBi
An: edsorb.342181
RelevancyScore: 1114
AccessLevel: 3
PubType: Conference
PubTypeId: conference
PreciseRelevancyScore: 1113.62585449219
IllustrationInfo
Items – Name: Title
  Label: Title
  Group: Ti
  Data: Appa: Bending Weather Dynamics with Latent Diffusion Models for Global Data Assimilation
– Name: Author
  Label: Authors
  Group: Au
  Data: <searchLink fieldCode="AR" term="%22Andry%2C+Gérôme%22">Andry, Gérôme</searchLink><br /><searchLink fieldCode="AR" term="%22Lewin%2C+Sacha%22">Lewin, Sacha</searchLink><br /><searchLink fieldCode="AR" term="%22Rozet%2C+François%22">Rozet, François</searchLink><br /><searchLink fieldCode="AR" term="%22Rochman%2C+Omer%22">Rochman, Omer</searchLink><br /><searchLink fieldCode="AR" term="%22Mangeleer%2C+Victor%22">Mangeleer, Victor</searchLink><br /><searchLink fieldCode="AR" term="%22Pirlet%2C+Matthias%22">Pirlet, Matthias</searchLink><br /><searchLink fieldCode="AR" term="%22Faulx%2C+Elise%22">Faulx, Elise</searchLink><br /><searchLink fieldCode="AR" term="%22Grégoire%2C+Marilaure%22">Grégoire, Marilaure</searchLink><br /><searchLink fieldCode="AR" term="%22Louppe%2C+Gilles%22">Louppe, Gilles</searchLink>
– Name: TitleSource
  Label: Source
  Group: Src
  Data: Machine Learning and the Physical Sciences Workshop (NeurIPS 2025), San Diego, United States - California [US-CA], 06/12/2025
– Name: DatePubCY
  Label: Publication Year
  Group: Date
  Data: 2025
– Name: Subject
  Label: Subject Terms
  Group: Su
  Data: <searchLink fieldCode="DE" term="%22Computer+Science+-+Learning%22">Computer Science - Learning</searchLink><br /><searchLink fieldCode="DE" term="%22Physics+-+Atmospheric+and+Oceanic+Physics%22">Physics - Atmospheric and Oceanic Physics</searchLink><br /><searchLink fieldCode="DE" term="%22Engineering%2C+computing+%26+technology%22">Engineering, computing & technology</searchLink><br /><searchLink fieldCode="DE" term="%22Computer+science%22">Computer science</searchLink><br /><searchLink fieldCode="DE" term="%22Ingénierie%2C+informatique+%26+technologie%22">Ingénierie, informatique & technologie</searchLink><br /><searchLink fieldCode="DE" term="%22Sciences+informatiques%22">Sciences informatiques</searchLink>
– Name: Abstract
  Label: Description
  Group: Ab
  Data: Deep learning has advanced weather forecasting, but accurate predictions first require identifying the current state of the atmosphere from observational data. In this work, we introduce Appa, a score-based data assimilation model generating global atmospheric trajectories at 0.25\si{\degree} resolution and 1-hour intervals. Powered by a 565M-parameter latent diffusion model trained on ERA5, Appa can be conditioned on arbitrary observations to infer plausible trajectories, without retraining. Our probabilistic framework handles reanalysis, filtering, and forecasting, within a single model, producing physically consistent reconstructions from various inputs. Results establish latent score-based data assimilation as a promising foundation for future global atmospheric modeling systems.
– Name: TypeDocument
  Label: Document Type
  Group: TypDoc
  Data: conference paper not in proceedings<br />http://purl.org/coar/resource_type/c_18cp<br />conferencePaper<br />peer reviewed
– Name: Language
  Label: Language
  Group: Lang
  Data: English
– Name: DOI
  Label: DOI
  Group: ID
  Data: 10.48550/arXiv.2504.18720
– Name: URL
  Label: Access URL
  Group: URL
  Data: <link linkTarget="URL" linkTerm="https://orbi.uliege.be/handle/2268/342181" linkWindow="_blank">https://orbi.uliege.be/handle/2268/342181</link>
– Name: Copyright
  Label: Rights
  Group: Cpyrght
  Data: open access<br />http://purl.org/coar/access_right/c_abf2<br />info:eu-repo/semantics/openAccess
– Name: AN
  Label: Accession Number
  Group: ID
  Data: edsorb.342181
PLink https://search.ebscohost.com/login.aspx?direct=true&site=eds-live&db=edsorb&AN=edsorb.342181
RecordInfo BibRecord:
  BibEntity:
    Identifiers:
      – Type: doi
        Value: 10.48550/arXiv.2504.18720
    Languages:
      – Text: English
    Subjects:
      – SubjectFull: Computer Science - Learning
        Type: general
      – SubjectFull: Physics - Atmospheric and Oceanic Physics
        Type: general
      – SubjectFull: Engineering, computing & technology
        Type: general
      – SubjectFull: Computer science
        Type: general
      – SubjectFull: Ingénierie, informatique & technologie
        Type: general
      – SubjectFull: Sciences informatiques
        Type: general
    Titles:
      – TitleFull: Appa: Bending Weather Dynamics with Latent Diffusion Models for Global Data Assimilation
        Type: main
  BibRelationships:
    HasContributorRelationships:
      – PersonEntity:
          Name:
            NameFull: Andry, Gérôme
      – PersonEntity:
          Name:
            NameFull: Lewin, Sacha
      – PersonEntity:
          Name:
            NameFull: Rozet, François
      – PersonEntity:
          Name:
            NameFull: Rochman, Omer
      – PersonEntity:
          Name:
            NameFull: Mangeleer, Victor
      – PersonEntity:
          Name:
            NameFull: Pirlet, Matthias
      – PersonEntity:
          Name:
            NameFull: Faulx, Elise
      – PersonEntity:
          Name:
            NameFull: Grégoire, Marilaure
      – PersonEntity:
          Name:
            NameFull: Louppe, Gilles
    IsPartOfRelationships:
      – BibEntity:
          Dates:
            – D: 06
              M: 12
              Type: published
              Y: 2025
ResultId 1