pith. machine review for the scientific record. sign in

arxiv: 2605.06337 · v1 · submitted 2026-05-07 · 💻 cs.CV

Recognition: unknown

Earth-o1: A Grid-free Observation-native Atmospheric World Model

Junchao Gong , Kaiyi Xu , Wangxu Wei , Siwei Tu , Jingyi Xu , Zili Liu , Hang Fan , Zhiwang Zhou
show 17 more authors
Tao Han Yi Xiao Xinyu Gu Zhangrui Li Wenlong Zhang Hao Chen Xiaokang Yang Yaqiang Wang Lijing Cheng Pierre Gentine Wanli Ouyang Feng Zhang Zhe-Min Tan Bowen Zhou Fenghua Ling Ben Fei Lei Bai
Authors on Pith no claims yet

Pith reviewed 2026-05-08 13:29 UTC · model grok-4.3

classification 💻 cs.CV
keywords atmospheric modelinggrid-freeobservation-nativeneural world modeldata-driven forecasthindcast evaluationEarth observation
0
0 comments X

The pith

Earth-o1 shows a grid-free neural model can advance continuous atmospheric states from raw observations and reach IFS-level surface forecast skill.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper presents Earth-o1 as an atmospheric simulator that processes raw Earth observation data without mapping it to fixed spatial grids. The model learns to maintain and advance a continuous three-dimensional representation of the atmosphere directly from heterogeneous sensor measurements. This removes the usual steps of data assimilation and numerical integration. Hindcast tests indicate that its surface predictions match the accuracy of the operational Integrated Forecasting System. The approach suggests a route to simulators that scale naturally with growing volumes of multimodal observations.

Core claim

Earth-o1 is an observation-native atmospheric world model that learns the continuous, three-dimensional physical evolution of the Earth system from ungridded observational data. It integrates diverse sensor inputs into a unified dynamical field and autonomously advances the atmospheric state in space and time without explicit numerical solvers or grid-based assimilation. Hindcast evaluations establish that this yields surface forecast skill comparable to the operational Integrated Forecasting System.

What carries the argument

The unified grid-free dynamical field that absorbs raw sensor data and evolves the atmospheric state forward continuously in space and time.

If this is right

  • Real-time forecasting becomes possible directly from incoming sensor streams without intermediate gridding.
  • Cross-sensor inference can fill observational gaps using the learned continuous field.
  • The framework supplies a scalable data-driven base for constructing a full digital twin of the Earth.
  • Computational cost drops because grid conversion and assimilation steps are eliminated.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The same continuous representation could support rapid updates whenever new observations arrive from any sensor type.
  • Hybrid versions might add selective physical constraints during training to improve behavior during extreme events.
  • Generalization to multi-year climate runs would test whether the learned dynamics remain consistent beyond short forecasts.

Load-bearing premise

A neural network can preserve physical consistency in the continuous three-dimensional atmospheric state when trained only on ungridded observations without any embedded numerical solvers or physics rules.

What would settle it

Independent hindcast runs over multiple seasons where Earth-o1 surface variable predictions are scored against IFS using standard metrics such as root-mean-square error and anomaly correlation coefficient.

read the original abstract

Despite the unprecedented volume of multimodal data provided by modern Earth observation systems, our ability to model atmospheric dynamics remains constrained. Traditional modeling frameworks force heterogeneous measurements into predefined spatial grids, inherently limiting the full exploitation of raw sensor data and creating severe computational bottlenecks. Here we present Earth-o1, an observation-native atmospheric world model that overcomes these structural limitations. Rather than relying on conventional atmospheric dynamical modeling systems or traditional data assimilation, Earth-o1 directly learns the continuous, three-dimensional physical evolution of the Earth system from ungridded observational data. By integrating diverse sensor inputs into a unified, grid-free dynamical field, the model autonomously advances the atmospheric state in space and time. We show that this fundamentally distinct paradigm enables direct, real-time forecasting and cross-sensor inference without the overhead of explicit numerical solvers. In hindcast evaluations, Earth-o1 achieves surface forecast skill comparable to the operational Integrated Forecasting System (IFS). These results establish that continuous, observation-driven world models -- a new class of fully observation-native geophysical simulators -- can match the fidelity of established physical frameworks, providing a scalable data-driven foundation for a digital twin of the Earth.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

3 major / 1 minor

Summary. The manuscript introduces Earth-o1, a grid-free observation-native atmospheric world model that directly learns the continuous three-dimensional physical evolution of the Earth system from ungridded observational data. It claims to autonomously advance the atmospheric state in space and time without explicit numerical solvers or grid-based assimilation, enabling real-time forecasting and cross-sensor inference, and reports that in hindcast evaluations Earth-o1 achieves surface forecast skill comparable to the operational Integrated Forecasting System (IFS).

Significance. If the central claims hold after detailed validation, this would represent a notable shift toward fully data-driven continuous geophysical simulators that bypass traditional discretization and assimilation pipelines. The potential to directly ingest raw multimodal sensor data at scale could reduce computational overhead and support new digital-twin applications, though the absence of supporting evidence currently limits assessment of whether the approach genuinely matches or exceeds established physical frameworks.

major comments (3)
  1. Abstract: the claim that 'Earth-o1 achieves surface forecast skill comparable to the operational Integrated Forecasting System (IFS)' supplies no evaluation metrics, baseline comparisons, error bars, data-handling procedures, or hindcast period details, rendering the central empirical claim impossible to assess.
  2. Model and Results sections: no quantitative checks on physical consistency (conservation of mass, momentum, energy, or moisture) or long-term stability are reported for the grid-free continuous-state advancement, which is load-bearing for the assertion that the model can autonomously evolve the 3D atmospheric state without explicit solvers.
  3. Training and Architecture description: the manuscript provides no details on network architecture, loss functions, training data volume or preprocessing, or how physical constraints are implicitly learned from ungridded observations, preventing evaluation of whether the reported skill arises from genuine dynamical modeling or from hindcast overfitting.
minor comments (1)
  1. Notation for the 'unified, grid-free dynamical field' is introduced without a formal definition or equation, making it difficult to distinguish the representation from standard latent-space embeddings.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for their constructive and detailed review. The comments identify important gaps in the presentation of empirical evidence, physical validation, and methodological transparency. We will undertake a major revision to address each point by adding the requested quantitative details, consistency checks, and architectural specifications. These changes will allow a fuller evaluation of the claims without altering the core contributions of the work.

read point-by-point responses

  1. Referee: Abstract: the claim that 'Earth-o1 achieves surface forecast skill comparable to the operational Integrated Forecasting System (IFS)' supplies no evaluation metrics, baseline comparisons, error bars, data-handling procedures, or hindcast period details, rendering the central empirical claim impossible to assess.

    Authors: We agree that the abstract presents the skill claim without supporting specifics, which limits immediate assessment. In the revised manuscript we will expand the abstract to include key quantitative metrics (e.g., RMSE and anomaly correlation coefficients for surface variables), the hindcast evaluation period, explicit baseline comparisons with IFS, and references to error bars and data-handling procedures that are described in the Methods and Results sections. This will make the central claim verifiable while respecting abstract length limits. revision: yes

  2. Referee: Model and Results sections: no quantitative checks on physical consistency (conservation of mass, momentum, energy, or moisture) or long-term stability are reported for the grid-free continuous-state advancement, which is load-bearing for the assertion that the model can autonomously evolve the 3D atmospheric state without explicit solvers.

    Authors: We acknowledge that explicit quantitative checks on physical consistency and long-term stability are essential to support the claim of autonomous, solver-free evolution. The current manuscript emphasizes forecast skill but does not report these diagnostics. In the revision we will add dedicated analyses in the Results section, including integrated budgets for mass, energy, and moisture over forecast horizons, as well as extended multi-day rollout experiments demonstrating stability without divergence. These checks will be adapted to the continuous, observation-native representation. revision: yes

  3. Referee: Training and Architecture description: the manuscript provides no details on network architecture, loss functions, training data volume or preprocessing, or how physical constraints are implicitly learned from ungridded observations, preventing evaluation of whether the reported skill arises from genuine dynamical modeling or from hindcast overfitting.

    Authors: We agree that the absence of these details prevents proper evaluation of reproducibility and the source of the reported skill. The revised manuscript will substantially expand the Training and Architecture sections to specify the network architecture (layer types, dimensions, and connectivity), the full loss function formulation, the volume and temporal/spatial coverage of training observations, preprocessing steps applied to ungridded sensor data, and an analysis of how the learned representation captures dynamical structure. We will also include discussion of validation strategies used to address potential overfitting. revision: yes

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Abstract provides no explicit free parameters, axioms, or invented entities; full text would be required to audit any learned components or physical assumptions.

pith-pipeline@v0.9.0 · 5578 in / 1001 out tokens · 51140 ms · 2026-05-08T13:29:38.050198+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

44 extracted references · 14 canonical work pages · 5 internal anchors

  1. [1]

    Nature641(8065), 1172–1179 (2025)

    Allen, A., Markou, S., Tebbutt, W., Requeima, J., Bruinsma, W.P., Andersson, T.R., Herzog, M., Lane, N.D., Chantry, M., Hosking, J.S.,et al.: End-to-end data-driven weather prediction. Nature641(8065), 1172–1179 (2025)

    2025

  2. [2]

    Anna Allen, Stratis Markou, Will Tebbutt, James Requeima, Wessel P Bruinsma, Tom R Andersson, Michael Herzog, Nicholas D Lane, Matthew Chantry, J Scott Hosking, et al

    Alexe, M., Boucher, E., Lean, P., Pinnington, E., Laloyaux, P., McNally, A., Lang, S., Chantry, M., Burrows, C., Chrust, M., et al.: Graphdop: Towards skilful data-driven medium-range weather forecasts learnt and initialised directly from observations. arXiv preprint arXiv:2412.15687 (2024)

    work page arXiv 2024

  3. [3]

    Feng, L., Li, D.: Can satellites and ai reveal the maxwell equations of the earth? National Science Review, 210 (2026)

    2026

  4. [4]

    Bulletin of the American Meteorological Society 98(8), 1637–1658 (2017)

    Yang, J., Zhang, Z., Wei, C., Lu, F., Guo, Q.: Introducing the new generation of chinese geostationary weather satellites, fengyun-4. Bulletin of the American Meteorological Society 98(8), 1637–1658 (2017)

    2017

  5. [5]

    Journal of the Meteorological Society of Japan

    Bessho, K., Date, K., Hayashi, M., Ikeda, A., Imai, T., Inoue, H., Kumagai, Y., Miyakawa, 21 T., Murata, H., Ohno, T.,et al.: An introduction to himawari-8/9—japan’s new-generation geostationary meteorological satellites. Journal of the Meteorological Society of Japan. Ser. II 94(2), 151–183 (2016)

    2016

  6. [6]

    Astroparticle Physics35(9), 591–607 (2012)

    Abreu, P., Aglietta, M., Ahlers, M., Ahn, E., Albuquerque, I.F.d.M., Allard, D., Allekotte, I., Allen, J., Allison, P., Almela, A.,et al.: Description of atmospheric conditions at the pierre auger observatory using the global data assimilation system (gdas). Astroparticle Physics35(9), 591–607 (2012)

    2012

  7. [7]

    Nature619(7970), 526–532 (2023)

    Zhang, Y., Long, M., Chen, K., Xing, L., Jin, R., Jordan, M.I., Wang, J.: Skilful nowcasting of extreme precipitation with nowcastnet. Nature619(7970), 526–532 (2023)

    2023

  8. [8]

    Nature597(7878), 672–677 (2021)

    Ravuri, S., Lenc, K., Willson, M., Kangin, D., Lam, R., Mirowski, P., Fitzsimons, M., Athanassiadou, M., Kashem, S., Madge, S.,et al.: Skilful precipitation nowcasting using deep generative models of radar. Nature597(7878), 672–677 (2021)

    2021

  9. [9]

    Nature619(7970), 533–538 (2023)

    Bi, K., Xie, L., Zhang, H., Chen, X., Gu, X., Tian, Q.: Accurate medium-range global weather forecasting with 3d neural networks. Nature619(7970), 533–538 (2023)

    2023

  10. [10]

    Science382(6677), 1416–1421 (2023)

    Lam, R., Sanchez-Gonzalez, A., Willson, M., Wirnsberger, P., Fortunato, M., Alet, F., Ravuri, S., Ewalds, T., Eaton-Rosen, Z., Hu, W.,et al.: Learning skillful medium-range global weather forecasting. Science382(6677), 1416–1421 (2023)

    2023

  11. [11]

    Fengwu: Pushing the skillful global medium-range weather forecast beyond 10 days lead,

    Chen, K., Han, T., Gong, J., Bai, L., Ling, F., Luo, J.-J., Chen, X., Ma, L., Zhang, T., Su, R., et al.: Fengwu: Pushing the skillful global medium-range weather forecast beyond 10 days lead. arXiv preprint arXiv:2304.02948 (2023)

    work page arXiv 2023

  12. [12]

    arXiv preprint arXiv:2402.00059 (2024)

    Han, T., Guo, S., Ling, F., Chen, K., Gong, J., Luo, J., Gu, J., Dai, K., Ouyang, W., Bai, L.: Fengwu-ghr: Learning the kilometer-scale medium-range global weather forecasting. arXiv preprint arXiv:2402.00059 (2024)

    work page arXiv 2024

  13. [13]

    arXiv p reprint arXiv:2407.15586 (2024)

    McNally, A., Lessig, C., Lean, P., Boucher, E., Alexe, M., Pinnington, E., Chantry, M., Lang, S., Burrows, C., Chrust, M., et al.: Data driven weather forecasts trained and initialised directly from observations. arXiv preprint arXiv:2407.15586 (2024)

    work page arXiv 2024

  14. [14]

    i: Experimental results with simplified physics

    Rabier, F., J¨ arvinen, H., Klinker, E., Mahfouf, J.-F., Simmons, A.: The ecmwf operational implementation of four-dimensional variational assimilation. i: Experimental results with simplified physics. Quarterly Journal of the Royal Meteorological Society126(564), 1143–1170 (2000)

    2000

  15. [15]

    Quarterly journal of the 22 royal meteorological society146(730), 1999–2049 (2020)

    Hersbach, H., Bell, B., Berrisford, P., Hirahara, S., Hor´ anyi, A., Mu˜ noz-Sabater, J., Nicolas, J., Peubey, C., Radu, R., Schepers, D.,et al.: The era5 global reanalysis. Quarterly journal of the 22 royal meteorological society146(730), 1999–2049 (2020)

    1999

  16. [16]

    arXiv preprint arXiv:2411.10191 (2024)

    Ling, F., Chen, K., Wu, J., Han, T., Luo, J.-J., Ouyang, W., Bai, L.: Fengwu-w2s: A deep learning model for seamless weather-to-subseasonal forecast of global atmosphere. arXiv preprint arXiv:2411.10191 (2024)

    work page arXiv 2024

  17. [17]

    Science Advances11(33), 2488 (2025)

    Guo, Z., Lyu, P., Ling, F., Bai, L., Luo, J.-J., Boers, N., Yamagata, T., Izumo, T., Cravatte, S., Capotondi, A.,et al.: Data-driven global ocean modeling for seasonal to decadal prediction. Science Advances11(33), 2488 (2025)

    2025

  18. [18]

    arXiv preprint arXiv:2312.12455 (2023)

    Xiao, Y., Bai, L., Xue, W., Chen, K., Han, T., Ouyang, W.: Fengwu-4dvar: Coupling the data-driven weather forecasting model with 4d variational assimilation. arXiv preprint arXiv:2312.12455 (2023)

    work page arXiv 2023

  19. [19]

    arXiv preprint arXiv:2312.12462 (2023)

    Chen, K., Bai, L., Ling, F., Ye, P., Chen, T., Luo, J.-J., Chen, H., Xiao, Y., Chen, K., Han, T., et al.: Towards an end-to-end artificial intelligence driven global weather forecasting system. arXiv preprint arXiv:2312.12462 (2023)

    work page arXiv 2023

  20. [20]

    Advances in Neural Information Processing Systems37, 137847–137872 (2024)

    Chen, K., Ye, P., Chen, H., Chen, K., Han, T., Ouyang, W., Chen, T., Bai, L.: Fnp: Fourier neural processes for arbitrary-resolution data assimilation. Advances in Neural Information Processing Systems37, 137847–137872 (2024)

    2024

  21. [21]

    Cosmos World Foundation Model Platform for Physical AI

    Agarwal, N., Ali, A., Bala, M., Balaji, Y., Barker, E., Cai, T., Chattopadhyay, P., Chen, Y., Cui, Y., Ding, Y., et al.: Cosmos world foundation model platform for physical ai. arXiv preprint arXiv:2501.03575 (2025)

    work page internal anchor Pith review arXiv 2025

  22. [22]

    Wedi, N., Bauer, P., Denoninck, W., Diamantakis, M., Hamrud, M., Kuhnlein, C., Malardel, S., Mogensen, K., Mozdzynski, G., Smolarkiewicz, P.: The modelling infrastructure of the integrated forecasting system: Recent advances and future challenges (2015)

    2015

  23. [23]

    In: 2025 IEEE International Conference on Data Mining Workshops (ICDMW), pp

    Vandal, T.J., Duffy, K., Nachmany, Y., McDuff, D.: Global atmospheric data assimilation with multi-modal masked autoencoders. In: 2025 IEEE International Conference on Data Mining Workshops (ICDMW), pp. 863–872 (2025). IEEE

    2025

  24. [24]

    Advances in Neural Information Processing Systems36, 58363– 58408 (2023)

    Mizrahi, D., Bachmann, R., Kar, O., Yeo, T., Gao, M., Dehghan, A., Zamir, A.: 4m: Massively multimodal masked modeling. Advances in Neural Information Processing Systems36, 58363– 58408 (2023)

    2023

  25. [25]

    5 technical report

    Yang, A., Yang, B., Zhang, B., Hui, B., Zheng, B., Yu, B., Li, C., Liu, D., Huang, F., Wei, H., et al.: Qwen2. 5 technical report. arXiv e-prints, 2412 (2024)

    2024

  26. [26]

    5-vl technical report

    Bai, S., Chen, K., Liu, X., Wang, J., Ge, W., Song, S., Dang, K., Wang, P., Wang, S., Tang, J., et al.: Qwen2. 5-vl technical report. arXiv e-prints, 2502 (2025) 23

    2025

  27. [27]

    Qwen3 Technical Report

    Yang, A., Li, A., Yang, B., Zhang, B., Hui, B., Zheng, B., Yu, B., Gao, C., Huang, C., Lv, C., et al.: Qwen3 technical report. arXiv preprint arXiv:2505.09388 (2025)

    work page internal anchor Pith review arXiv 2025

  28. [28]

    Springer (2021)

    Zhang, Q., Ng, C.-P., Dai, K., Xu, J., Tang, J., Sun, J., Mu, M.: Lessons learned from the tragedy during the 100 km ultramarathon race in Baiyin, Gansu Province on 22 May 2021. Springer (2021)

    2021

  29. [29]

    IEEE Transactions on Geoscience and Remote Sensing53(4), 1985–1996 (2014)

    Beitsch, A., Kern, S., Kaleschke, L.: Comparison of ssm/i and amsr-e sea ice concentrations with aspect ship observations around antarctica. IEEE Transactions on Geoscience and Remote Sensing53(4), 1985–1996 (2014)

    1985

  30. [30]

    Geophysical Research Letters43(3), 1234–1240 (2016)

    Kurihara, Y., Murakami, H., Kachi, M.: Sea surface temperature from the new japanese geostationary meteorological himawari-8 satellite. Geophysical Research Letters43(3), 1234–1240 (2016)

    2016

  31. [31]

    IEEE Geoscience and Remote Sensing Letters7(3), 440–444 (2010)

    Surussavadee, C., Staelin, D.H.: Npoess precipitation retrievals using the atms passive microwave spectrometer. IEEE Geoscience and Remote Sensing Letters7(3), 440–444 (2010)

    2010

  32. [32]

    Atmospheric Measurement Techniques13(8), 4437–4459 (2020)

    Smith, N., Barnet, C.D.: Climcaps observing capability for temperature, moisture, and trace gases from airs/amsu and cris/atms. Atmospheric Measurement Techniques13(8), 4437–4459 (2020)

    2020

  33. [33]

    IEEE Transactions on Geoscience and Remote Sensing60, 1–11 (2022)

    Ri, X., Tana, G., Shi, C., Nakajima, T.Y., Shi, J., Zhao, J., Xu, J., Letu, H.: Cloud, atmospheric radiation and renewal energy application (care) version 1.0 cloud top property product from himawari-8/ahi: Algorithm development and preliminary validation. IEEE Transactions on Geoscience and Remote Sensing60, 1–11 (2022)

    2022

  34. [34]

    Meteorological Satellite Center Technical Note65, 42 (2020)

    Shimizu, A.: Introduction to himawari-8 rgb composite imagery. Meteorological Satellite Center Technical Note65, 42 (2020)

    2020

  35. [35]

    Scaling Laws for Neural Language Models

    Kaplan, J., McCandlish, S., Henighan, T., Brown, T.B., Chess, B., Child, R., Gray, S., Radford, A., Wu, J., Amodei, D.: Scaling laws for neural language models. arXiv preprint arXiv:2001.08361 (2020)

    work page internal anchor Pith review arXiv 2001

  36. [36]

    Training Compute-Optimal Large Language Models

    Hoffmann, J., Borgeaud, S., Mensch, A., Buchatskaya, E., Cai, T., Rutherford, E., Casas, D., Hendricks, L.A., Welbl, J., Clark, A., et al.: Training compute-optimal large language models. arXiv preprint arXiv:2203.1555610(2022)

    work page internal anchor Pith review arXiv 2022

  37. [37]

    Nature637(8044), 84–90 (2025)

    Price, I., Sanchez-Gonzalez, A., Alet, F., Andersson, T.R., El-Kadi, A., Masters, D., Ewalds, T., Stott, J., Mohamed, S., Battaglia, P.,et al.: Probabilistic weather forecasting with machine learning. Nature637(8044), 84–90 (2025)

    2025

  38. [38]

    24 Communications Earth & Environment6(1), 955 (2025)

    Han, T., Chen, Z., Guo, S., Xu, W., Ouyang, W., Bai, L.: Climate science data can be compressed efficiently by dual-stage extreme compression with a variational auto-encoder transformer. 24 Communications Earth & Environment6(1), 955 (2025)

    2025

  39. [39]

    IEEE Transactions on Pattern Analysis and Machine Intelligence47(3), 2020–2036 (2024)

    Li, Z., Wang, W., Li, H., Xie, E., Sima, C., Lu, T., Yu, Q., Dai, J.: Bevformer: learning bird’s- eye-view representation from lidar-camera via spatiotemporal transformers. IEEE Transactions on Pattern Analysis and Machine Intelligence47(3), 2020–2036 (2024)

    2020

  40. [40]

    In: Proceedings of the IEEE/CVF International Conference on Computer Vision, pp

    Zhao, H., Jiang, L., Jia, J., Torr, P.H., Koltun, V.: Point transformer. In: Proceedings of the IEEE/CVF International Conference on Computer Vision, pp. 16259–16268 (2021)

    2021

  41. [41]

    In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp

    Rombach, R., Blattmann, A., Lorenz, D., Esser, P., Ommer, B.: High-resolution image synthesis with latent diffusion models. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 10684–10695 (2022)

    2022

  42. [42]

    An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale

    Dosovitskiy, A., Beyer, L., Kolesnikov, A., Weissenborn, D., Zhai, X., Unterthiner, T., Dehghani, M., Minderer, M., Heigold, G., Gelly, S., et al.: An image is worth 16x16 words: Transformers for image recognition at scale. arXiv preprint arXiv:2010.11929 (2020)

    work page internal anchor Pith review arXiv 2010

  43. [43]

    Neural processes,

    Garnelo, M., Schwarz, J., Rosenbaum, D., Viola, F., Rezende, D.J., Eslami, S., Teh, Y.W.: Neural processes. arXiv preprint arXiv:1807.01622 (2018)

    work page arXiv 2018

  44. [44]

    arXiv preprint arXiv:2510.15978 (2025) 25

    Gong, J., Xu, J., Fei, B., Ling, F., Zhang, W., Chen, K., Xu, W., Yang, W., Yang, X., Bai, L.: Dawp: A framework for global observation forecasting via data assimilation and weather prediction in satellite observation space. arXiv preprint arXiv:2510.15978 (2025) 25

    work page arXiv 2025