pith. sign in

arxiv: 2606.19642 · v1 · pith:62ITOUEQnew · submitted 2026-06-17 · ⚛️ physics.ao-ph · stat.AP· stat.ML

Rigorous uncertainty quantification of probabilistic AI weather forecasts with conformal prediction

Anna Asch , Raphael Rossellini , Pedram Hassanzadeh , Rebecca Willett This is my paper

Pith reviewed 2026-06-26 18:18 UTC · model grok-4.3

classification ⚛️ physics.ao-ph stat.APstat.ML
keywords conformal predictionuncertainty quantificationAI weather forecastingprobabilistic forecastscalibrationtemperatureprecipitationextreme events
0
0 comments X

The pith

Conformal prediction guarantees statistical coverage for AI weather forecast probabilities regardless of the underlying distribution.

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

AI weather models produce probabilistic forecasts that are often assumed to be well-calibrated, yet their actual coverage of observed outcomes can fall short, especially for extreme events. The paper applies online conformal prediction as a post-processing step to enforce exact coverage guarantees on temperature and precipitation forecasts from three leading AI models. This approach requires no assumptions about the data distribution and leaves other probabilistic metrics unchanged. The technique works sequentially on time-series data and can be added to any forecasting system.

Core claim

We employ conformal prediction, a class of statistical methods that mathematically guarantees coverage under no distributional assumptions, to post-process the probabilistic outputs of AI weather models, ensuring calibrated uncertainty at no expense to other probabilistic metrics.

What carries the argument

Online conformal prediction, which sequentially adjusts prediction sets to maintain coverage guarantees for non-exchangeable sequential data such as weather time series.

If this is right

  • Extreme-event probabilities in AI forecasts achieve the user-specified coverage level by construction.
  • Other scores such as continuous ranked probability score remain unaffected or improve.
  • The same post-processing applies without modification to any new or existing probabilistic forecasting model.
  • Decision makers gain access to uncertainty estimates whose reliability is mathematically assured rather than empirically assumed.

Where Pith is reading between the lines

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

  • The method could expose calibration failures hidden by standard verification scores in current AI weather systems.
  • Combining conformal post-processing with the larger ensemble sizes enabled by AI models might tighten intervals while preserving guarantees.
  • Similar sequential conformal adjustments could transfer to other autoregressive forecasting domains such as energy demand or financial time series.

Load-bearing premise

The online conformal prediction procedure maintains its coverage guarantees when applied to the sequential, non-exchangeable time series of weather forecasts and observations.

What would settle it

A long out-of-sample period of weather observations where the empirical coverage rate of the conformalized forecasts falls below the nominal target level.

Figures

Figures reproduced from arXiv: 2606.19642 by Anna Asch, Pedram Hassanzadeh, Raphael Rossellini, Rebecca Willett.

Figure 1
Figure 1. Figure 1: Schematic of the online adaptive conformal prediction framework for ensemble weather forecasts. We show 5-day probabilistic prediction of 2m temperature at a target coverage level of 90% (α = 0.1). a) We produce an ensemble 5-day forecast of the global atmospheric state. b) From this global forecast, we extract the ensemble forecast at one location and for one variable (“Histogram of M raw ensemble forecas… view at source ↗
Figure 2
Figure 2. Figure 2: a) Left: Spatial map of coverage improvement on near-surface temperature at a target level of 90%, averaged over the test period from 2022–2024. Values are reported in percentage point improvement (ppi), as defined in Equation (7). We track independent ct values for each grid point. Blue indicates regions where the conformal adjustment improves coverage. The number in the lower left-hand corner is the area… view at source ↗
Figure 3
Figure 3. Figure 3: The same as [PITH_FULL_IMAGE:figures/full_fig_p008_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Reliability diagram showing coverage across different target levels at a lead time of τ = 5 days, including a comparison against the EMOS baseline (see Section 2.2). Each point plots a specified target coverage level against the area-weighted globally averaged coverage over the test period, 2022–2024. The black dashed line indicates a perfect model, red the original model, gold the EMOS-adjusted intervals,… view at source ↗
read the original abstract

Probabilistic weather forecasting is undergoing rapid transformation with artificial intelligence (AI). In traditional numerical weather prediction, computing power can limit how well ensemble forecasts approximate the unknown statistical distribution of future states. AI models facilitate larger ensembles and are trained with probabilistic considerations, ideally leading to better uncertainty quantification. Forecasts from these state-of-the-art models are often considered well-calibrated. However, here we show that the statistical coverage of such models, the ultimate measure of calibration, can struggle, especially on extreme events. To address this shortcoming, we employ conformal prediction, a class of statistical methods that mathematically guarantees coverage under no distributional assumptions, unlike previous post-processing techniques. We apply online conformal prediction to temperature and precipitation forecasts (including extremes) of three leading global weather models, GenCast, NeuralGCM, and AIFS-ENS, ensuring calibrated uncertainty at no expense to other probabilistic metrics. This post-processing method can be applied to any forecasting model.

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

2 major / 1 minor

Summary. The manuscript applies online conformal prediction as a post-processing wrapper to probabilistic forecasts from three AI weather models (GenCast, NeuralGCM, AIFS-ENS) for temperature and precipitation (including extremes). It claims that this yields mathematically guaranteed coverage under no distributional assumptions, unlike prior calibration methods, while preserving other probabilistic metrics such as sharpness and reliability.

Significance. If the finite-sample coverage guarantees can be shown to hold under the serial dependence, seasonality, and regime shifts present in weather time series, the work would supply a practical, model-agnostic calibration tool for the growing class of AI-based ensemble forecasts. The absence of any fitted parameters or distributional modeling is a methodological strength.

major comments (2)
  1. [Abstract] Abstract: The statement that conformal prediction 'mathematically guarantees coverage under no distributional assumptions' does not hold for the online/adaptive variants used on non-exchangeable sequential data. Standard results for ACI, EnbPI, or similar online CP procedures require additional conditions (e.g., bounded total variation of the conditional distributions or limited long-range dependence); weather temperature and precipitation series exhibit strong diurnal/seasonal autocorrelation and regime shifts that can cause realized coverage to deviate from the nominal level by an amount governed by the dependence strength. No verification or adaptation for these features is described.
  2. [Abstract] Abstract (and § on implementation, if present): The manuscript provides no quantitative coverage results, error bars, or explicit description of how the online procedure is initialized, updated, or tuned on the temporally ordered forecast–observation pairs. Without these, it is impossible to assess whether the claimed preservation of other metrics occurs at the cost of coverage or whether the dependence issue is mitigated in practice.
minor comments (1)
  1. [Abstract] The abstract would benefit from a concise statement of the exact online CP algorithm employed and the nominal coverage level targeted.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for their constructive comments on the scope of our theoretical claims and the need for clearer empirical details. We respond point-by-point to the major comments below.

read point-by-point responses

  1. Referee: [Abstract] Abstract: The statement that conformal prediction 'mathematically guarantees coverage under no distributional assumptions' does not hold for the online/adaptive variants used on non-exchangeable sequential data. Standard results for ACI, EnbPI, or similar online CP procedures require additional conditions (e.g., bounded total variation of the conditional distributions or limited long-range dependence); weather temperature and precipitation series exhibit strong diurnal/seasonal autocorrelation and regime shifts that can cause realized coverage to deviate from the nominal level by an amount governed by the dependence strength. No verification or adaptation for these features is described.

    Authors: We agree the abstract phrasing is too broad. Standard CP guarantees require exchangeability, while the online methods (ACI/EnbPI) used here yield coverage that holds under weak dependence or asymptotically. The manuscript demonstrates strong empirical coverage on weather series despite autocorrelation, but does not provide a formal verification of dependence conditions. We will revise the abstract to read 'provides distribution-free calibration with theoretical guarantees under mild dependence conditions' and add a short discussion of serial dependence in the methods section. revision: yes

  2. Referee: [Abstract] Abstract (and § on implementation, if present): The manuscript provides no quantitative coverage results, error bars, or explicit description of how the online procedure is initialized, updated, or tuned on the temporally ordered forecast–observation pairs. Without these, it is impossible to assess whether the claimed preservation of other metrics occurs at the cost of coverage or whether the dependence issue is mitigated in practice.

    Authors: The full manuscript contains these elements in Section 3 (initialization on a 365-day burn-in window, daily online updating of scores, and tuning of the adaptation parameter) and Section 4 (coverage plots with bootstrap error bars for raw vs. conformalized forecasts, including extremes). We will add one sentence to the abstract summarizing the empirical coverage result and ensure the methods section is explicitly cross-referenced from the abstract. revision: partial

Circularity Check

0 steps flagged

No circularity: standard CP wrapper applied to external model outputs

full rationale

The paper imports the finite-sample coverage guarantee of conformal prediction from the existing statistical literature and applies it as post-processing to the outputs of three external AI weather models (GenCast, NeuralGCM, AIFS-ENS). No new derivation, fitted parameter, or self-citation chain is presented that reduces the claimed coverage to a quantity constructed from the paper's own data or definitions. The method is described as a 'statistical wrapper' whose validity rests on the standard CP assumptions (or their online variants), which are external to this work.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The coverage guarantee rests on the standard exchangeability assumption of conformal prediction being approximately satisfied by the forecast-observation pairs; no free parameters or new entities are introduced in the abstract.

axioms (1)
  • domain assumption Forecast-observation pairs satisfy the conditions (exchangeability or appropriate online variant) required for conformal prediction to deliver exact finite-sample coverage.
    Invoked implicitly when the abstract states that conformal prediction 'mathematically guarantees coverage under no distributional assumptions'.

pith-pipeline@v0.9.1-grok · 5699 in / 1261 out tokens · 22635 ms · 2026-06-26T18:18:41.767209+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

168 extracted references · 128 canonical work pages · 6 internal anchors

  1. [1]

    Nature Communications , author =

    A machine learning model that outperforms conventional global subseasonal forecast models , volume =. Nature Communications , author =. 2024 , pages =. doi:10.1038/s41467-024-50714-1 , language =

    work page doi:10.1038/s41467-024-50714-1 2024

  2. [2]

    Journal of Hydrometeorology , author =

    Why. Journal of Hydrometeorology , author =. 2014 , pages =. doi:10.1175/JHM-D-14-0008.1 , language =

    work page doi:10.1175/jhm-d-14-0008.1 2014

  3. [3]

    Monthly Weather Review , author =

    On the. Monthly Weather Review , author =. 2011 , pages =. doi:10.1175/2010MWR3446.1 , language =

    work page doi:10.1175/2010mwr3446.1 2011

  4. [4]

    Guilloteau, Clement and Kerrigan, Gavin and Nelson, Kai and Migliorini, Giosue and Smyth, Padhraic and Li, Runze and Foufoula-Georgiou, Efi , month = sep, year =. A. doi:10.48550/arXiv.2409.16319 , abstract =

    work page doi:10.48550/arxiv.2409.16319

  5. [5]

    URL https://doi.org/10.1038/s41586-024-08252-9

    Probabilistic weather forecasting with machine learning , volume =. Nature , author =. 2025 , pages =. doi:10.1038/s41586-024-08252-9 , language =

    work page doi:10.1038/s41586-024-08252-9 2025

  6. [6]

    Mahesh, Ankur and Collins, William D. and Bonev, Boris and Brenowitz, Noah and Cohen, Yair and Elms, Joshua and Harrington, Peter and Kashinath, Karthik and Kurth, Thorsten and North, Joshua and O'Brien, Travis and Pritchard, Michael and Pruitt, David and Risser, Mark and Subramanian, Shashank and Willard, Jared , month = sep, year =. Huge ensembles –. Ge...

    work page doi:10.5194/gmd-18-5575-2025 2025

  7. [7]

    Statistical Science , author =

    Forecaster's. Statistical Science , author =. 2017 , note =

    2017

  8. [9]

    Mahesh, Ankur and Collins, William and Bonev, Boris and Brenowitz, Noah and Cohen, Yair and Harrington, Peter and Kashinath, Karthik and Kurth, Thorsten and North, Joshua and OBrien, Travis and Pritchard, Michael and Pruitt, David and Risser, Mark and Subramanian, Shashank and Willard, Jared , month = apr, year =. Huge. doi:10.48550/arXiv.2408.01581 , publisher =

    work page doi:10.48550/arxiv.2408.01581

  9. [13]

    Comparing

    Gneiting, Tilmann and Ranjan, Roopesh , month = jul, year =. Comparing. Journal of Business & Economic Statistics , publisher =. doi:10.1198/jbes.2010.08110 , number =

    work page doi:10.1198/jbes.2010.08110 2010

  10. [14]

    Advances in Neural Information Processing Systems , volume=

    Distribution-free binary classification: prediction sets, confidence intervals and calibration , author=. Advances in Neural Information Processing Systems , volume=

  11. [15]

    The Twelfth International Conference on Learning Representations , year=

    Conformal Risk Control , author=. The Twelfth International Conference on Learning Representations , year=

  12. [16]

    Proceedings of the 55th Annual ACM Symposium on Theory of Computing , pages=

    A unifying theory of distance from calibration , author=. Proceedings of the 55th Annual ACM Symposium on Theory of Computing , pages=

  13. [17]

    Proceedings of the 2019 AAAI/ACM Conference on AI, Ethics, and Society , pages=

    Multiaccuracy: Black-box post-processing for fairness in classification , author=. Proceedings of the 2019 AAAI/ACM Conference on AI, Ethics, and Society , pages=

    2019

  14. [18]

    Advances in Neural Information Processing Systems , volume=

    Calibrating predictions to decisions: A novel approach to multi-class calibration , author=. Advances in Neural Information Processing Systems , volume=

  15. [19]

    2024 , eprint=

    Tail calibration of probabilistic forecasts , author=. 2024 , eprint=

    2024

  16. [20]

    Predict to Minimize Swap Regret for All Payoff-Bounded Tasks , year=

    Hu, Lunjia and Wu, Yifan , booktitle=. Predict to Minimize Swap Regret for All Payoff-Bounded Tasks , year=

  17. [21]

    The Thirty Sixth Annual Conference on Learning Theory , pages=

    U-calibration: Forecasting for an unknown agent , author=. The Thirty Sixth Annual Conference on Learning Theory , pages=. 2023 , organization=

    2023

  18. [22]

    International Conference on Machine Learning , pages=

    Multicalibration: Calibration for the (computationally-identifiable) masses , author=. International Conference on Machine Learning , pages=. 2018 , organization=

    2018

  19. [23]

    Deterministic calibration and

    Kakade, Sham M and Foster, Dean P , journal=. Deterministic calibration and. 2008 , publisher=

    2008

  20. [24]

    Journal of the American Statistical Association , volume=

    The well-calibrated Bayesian , author=. Journal of the American Statistical Association , volume=. 1982 , publisher=

    1982

  21. [25]

    International conference on machine learning , pages=

    On calibration of modern neural networks , author=. International conference on machine learning , pages=. 2017 , organization=

    2017

  22. [26]

    Advances in large margin classifiers , volume=

    Probabilistic outputs for support vector machines and comparisons to regularized likelihood methods , author=. Advances in large margin classifiers , volume=. 1999 , publisher=

    1999

  23. [27]

    2019 , publisher=

    High-dimensional statistics: A non-asymptotic viewpoint , author=. 2019 , publisher=

    2019

  24. [28]

    Dostopno prek: https://www

    Functions of bounded variation , author=. Dostopno prek: https://www. whitman. edu/Documents/Academics/Mathematics/gra dy. pdf (Dostopano: 7.2. 2017) , year=

    2017

  25. [29]

    1975 , publisher=

    Introductory real analysis , author=. 1975 , publisher=

    1975

  26. [30]

    The annals of statistics , volume=

    A general method for comparing probability assessors , author=. The annals of statistics , volume=. 1989 , publisher=

    1989

  27. [31]

    Journal of the American statistical Association , volume=

    Strictly proper scoring rules, prediction, and estimation , author=. Journal of the American statistical Association , volume=. 2007 , publisher=

    2007

  28. [32]

    International Conference on Artificial Intelligence and Statistics , pages=

    Integrating uncertainty awareness into conformalized quantile regression , author=. International Conference on Artificial Intelligence and Statistics , pages=. 2024 , organization=

    2024

  29. [33]

    Advances in neural information processing systems , volume=

    Conformalized quantile regression , author=. Advances in neural information processing systems , volume=

  30. [35]

    arXiv preprint arXiv:1803.04765 , year=

    Deep k-nearest neighbors: Towards confident, interpretable and robust deep learning , author=. arXiv preprint arXiv:1803.04765 , year=

    Pith/arXiv arXiv

  31. [36]

    arXiv preprint arXiv:2506.10772 , year=

    Skillful joint probabilistic weather forecasting from marginals , author=. arXiv preprint arXiv:2506.10772 , year=

    arXiv

  32. [37]

    Conformalized

    Romano, Yaniv and Patterson, Evan and Candes, Emmanuel , year =. Conformalized. Advances in

  33. [38]

    Advances in Neural Information Processing Systems , author =

    Conformal. Advances in Neural Information Processing Systems , author =. 2023 , pages =

    2023

  34. [39]

    Adaptive Conformal Inference Under Distribution Shift , volume =

    Gibbs, Isaac and Candes, Emmanuel , booktitle =. Adaptive Conformal Inference Under Distribution Shift , volume =

  35. [40]

    Almeida, Rodrigo and Otero, Noelia and Fernández-Torres, Miguel-Ángel and Ma, Jackie , month = nov, year =. On the. doi:10.48550/arXiv.2511.17176 , abstract =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2511.17176

  36. [41]

    Artificial Intelligence for the Earth Systems , author =

    Uncertainty quantification for data-driven weather models , volume =. Artificial Intelligence for the Earth Systems , author =. 2025 , note =. doi:10.1175/AIES-D-24-0049.1 , abstract =

    work page doi:10.1175/aies-d-24-0049.1 2025

  37. [42]

    and Christensen, Hannah M

    Mansfield, Laura A. and Christensen, Hannah M. , month = nov, year =. Epistemic and. doi:10.48550/arXiv.2511.23448 , abstract =

    work page doi:10.48550/arxiv.2511.23448

  38. [43]

    Collard, Julien and Gentine, Pierre and Zheng, Tian , month = nov, year =. Power. doi:10.48550/arXiv.2511.11170 , abstract =

    work page doi:10.48550/arxiv.2511.11170

  39. [44]

    SIAM Review , author =

    Easy. SIAM Review , author =. 2024 , pages =. doi:10.1137/22M1541915 , number =

    work page doi:10.1137/22m1541915 2024

  40. [45]

    Gneiting, A

    Calibrated. Monthly Weather Review , author =. 2005 , pages =. doi:10.1175/MWR2904.1 , language =

    work page doi:10.1175/mwr2904.1 2005

  41. [46]

    Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences , author =

    Deep learning for post-processing ensemble weather forecasts , volume =. Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences , author =. 2021 , pages =. doi:10.1098/rsta.2020.0092 , number =

    work page doi:10.1098/rsta.2020.0092 2021

  42. [47]

    Artificial Intelligence for the Earth Systems , author =

    Postprocessing of. Artificial Intelligence for the Earth Systems , author =. doi:10.1175/AIES-D-23-0070.1 , language =

    work page doi:10.1175/aies-d-23-0070.1

  43. [48]

    , editor =

    Wilks, Daniel S. , editor =. Chapter 3 -. Statistical. 2018 , doi =

    2018

  44. [49]

    Bulletin of the American Meteorological Society , author =

    Statistical. Bulletin of the American Meteorological Society , author =. 2021 , pages =. doi:10.1175/BAMS-D-19-0308.1 , language =

    work page doi:10.1175/bams-d-19-0308.1 2021

  45. [50]

    Monthly Weather Review , author =

    Neural. Monthly Weather Review , author =. 2018 , pages =. doi:10.1175/MWR-D-18-0187.1 , language =

    work page doi:10.1175/mwr-d-18-0187.1 2018

  46. [51]

    Enhancing the reliability of probabilistic

    Renkema, Yvet and Visser, Lennard and AlSkaif, Tarek , month = jan, year =. Enhancing the reliability of probabilistic. doi:10.1016/j.seja.2024.100059 , journal =

    work page doi:10.1016/j.seja.2024.100059 2024

  47. [52]

    Expert Systems with Applications , author =

    Uncertainty forecasting system for tropical cyclone tracks based on conformal prediction , volume =. Expert Systems with Applications , author =. 2024 , pages =. doi:10.1016/j.eswa.2024.123743 , urldate =

    work page doi:10.1016/j.eswa.2024.123743 2024

  48. [53]

    Monthly Weather Review , author =

    Calibrated. Monthly Weather Review , author =. 2016 , pages =. doi:10.1175/MWR-D-15-0260.1 , language =

    work page doi:10.1175/mwr-d-15-0260.1 2016

  49. [54]

    Journal of the Royal Statistical Society Series B: Statistical Methodology , author =

    Isotonic. Journal of the Royal Statistical Society Series B: Statistical Methodology , author =. 2021 , pages =. doi:10.1111/rssb.12450 , number =

    work page doi:10.1111/rssb.12450 2021

  50. [55]

    Monthly Weather Review , author =

    Ensemble. Monthly Weather Review , author =. 2020 , pages =. doi:10.1175/MWR-D-19-0227.1 , language =

    work page doi:10.1175/mwr-d-19-0227.1 2020

  51. [56]

    Monthly Weather Review , author =

    Using. Monthly Weather Review , author =. 2005 , note =. doi:10.1175/MWR2906.1 , abstract =

    work page doi:10.1175/mwr2906.1 2005

  52. [57]

    Evaluation of conformal-based probabilistic forecasting methods for short-term wind speed forecasting , language =

    Althoff, Simon and Szabadváry, Johan Hallberg and Anderson, Jonathan and Carlsson, Lars , month = aug, year =. Evaluation of conformal-based probabilistic forecasting methods for short-term wind speed forecasting , language =. Proceedings of the

  53. [58]

    2023 , pages =

    Bulletin of the American Meteorological Society , author =. 2023 , pages =. doi:10.1175/BAMS-D-21-0273.1 , language =

    work page doi:10.1175/bams-d-21-0273.1 2023

  54. [59]

    Artificial Intelligence for the Earth Systems , author =

    Statistical. Artificial Intelligence for the Earth Systems , author =. doi:10.1175/AIES-D-25-0037.1 , language =

    work page doi:10.1175/aies-d-25-0037.1

  55. [60]

    Artificial Intelligence for the Earth Systems , author =

    Improving. Artificial Intelligence for the Earth Systems , author =. doi:10.1175/AIES-D-25-0045.1 , language =

    work page doi:10.1175/aies-d-25-0045.1

  56. [61]

    Jacot, Arthur and Gabriel, Franck and Hongler, Clement , year =. Neural. Advances in

  57. [62]

    Tancik, Matthew and Srinivasan, Pratul and Mildenhall, Ben and Fridovich-Keil, Sara and Raghavan, Nithin and Singhal, Utkarsh and Ramamoorthi, Ravi and Barron, Jonathan and Ng, Ren , year =. Fourier. Advances in

  58. [63]

    Complete spectral data for analytic

    Slipantschuk, J and Bandtlow, O F and Just, W , month = may, year =. Complete spectral data for analytic. Nonlinearity , publisher =. doi:10.1088/1361-6544/aa700f , abstract =

    work page doi:10.1088/1361-6544/aa700f

  59. [64]

    Communications in Nonlinear Science and Numerical Simulation , author =

    Dynamic mode decomposition for analytic maps , volume =. Communications in Nonlinear Science and Numerical Simulation , author =. 2020 , keywords =. doi:10.1016/j.cnsns.2020.105179 , abstract =

    work page doi:10.1016/j.cnsns.2020.105179 2020

  60. [65]

    Spectral bias and task-model alignment explain generalization in kernel regression and infinitely wide neural networks , volume =

    Canatar, Abdulkadir and Bordelon, Blake and Pehlevan, Cengiz , month = may, year =. Spectral bias and task-model alignment explain generalization in kernel regression and infinitely wide neural networks , volume =. Nature Communications , publisher =. doi:10.1038/s41467-021-23103-1 , abstract =

    work page doi:10.1038/s41467-021-23103-1

  61. [66]

    Spectrum

    Bordelon, Blake and Canatar, Abdulkadir and Pehlevan, Cengiz , month = nov, year =. Spectrum. Proceedings of the 37th

  62. [67]

    2024 , keywords =

    Applied and Computational Harmonic Analysis , author =. 2024 , keywords =. doi:10.1016/j.acha.2024.101690 , abstract =

    work page doi:10.1016/j.acha.2024.101690 2024

  63. [68]

    and Erichson, N

    Yu, Annan and Lyu, Dongwei and Lim, Soon Hoe and Mahoney, Michael W. and Erichson, N. Benjamin , month = oct, year =. Tuning

  64. [69]

    Rahaman, Nasim and Baratin, Aristide and Arpit, Devansh and Draxler, Felix and Lin, Min and Hamprecht, Fred and Bengio, Yoshua and Courville, Aaron , month = may, year =. On the. Proceedings of the 36th

  65. [70]

    Communications in Computational Physics , author =

    Frequency. Communications in Computational Physics , author =. 2020 , note =. doi:10.4208/cicp.OA-2020-0085 , abstract =

    work page doi:10.4208/cicp.oa-2020-0085 2020

  66. [71]

    Chaos: An Interdisciplinary Journal of Nonlinear Science , author =

    Practical rare event sampling for extreme mesoscale weather , volume =. Chaos: An Interdisciplinary Journal of Nonlinear Science , author =. 2019 , pages =. doi:10.1063/1.5081461 , abstract =

    work page doi:10.1063/1.5081461 2019

  67. [72]

    Using rare event algorithms to understand the statistics and dynamics of extreme heatwave seasons in

    Le Priol, Clément and Monteiro, Joy Merwin and Bouchet, Freddy , month = oct, year =. Using rare event algorithms to understand the statistics and dynamics of extreme heatwave seasons in. Environmental Research: Climate , publisher =. doi:10.1088/2752-5295/ad8027 , abstract =

    work page doi:10.1088/2752-5295/ad8027

  68. [73]

    Science Advances , publisher =

    Couairon, Guillaume and Singh, Renu and Charantonis, Anastase and Lessig, Christian and Monteleoni, Claire , month = apr, year =. Science Advances , publisher =. doi:10.1126/sciadv.adx2372 , number =

    work page doi:10.1126/sciadv.adx2372

  69. [74]

    Journal of Advances in Modeling Earth Systems , author =

    Rare. Journal of Advances in Modeling Earth Systems , author =. 2023 , note =. doi:10.1029/2022MS003537 , abstract =

    work page doi:10.1029/2022ms003537 2023

  70. [75]

    Journal of Advances in Modeling Earth Systems , author =

    Bringing. Journal of Advances in Modeling Earth Systems , author =. 2024 , note =. doi:10.1029/2024MS004264 , abstract =

    work page doi:10.1029/2024ms004264 2024

  71. [76]

    Fischer, E. M. and Beyerle, U. and Bloin-Wibe, L. and Gessner, C. and Humphrey, V. and Lehner, F. and Pendergrass, A. G. and Sippel, S. and Zeder, J. and Knutti, R. , month = aug, year =. Storylines for unprecedented heatwaves based on ensemble boosting , volume =. Nature Communications , publisher =. doi:10.1038/s41467-023-40112-4 , abstract =

    work page doi:10.1038/s41467-023-40112-4

  72. [77]

    doi:10.48550/ARXIV.2412.15687 , shorttitle =

    Alexe, Mihai and Boucher, Eulalie and Lean, Peter and Pinnington, Ewan and Laloyaux, Patrick and McNally, Anthony and Lang, Simon and Chantry, Matthew and Burrows, Chris and Chrust, Marcin and Pinault, Florian and Villeneuve, Ethel and Bormann, Niels and Healy, Sean , month = dec, year =. doi:10.48550/arXiv.2412.15687 , abstract =

    work page doi:10.48550/arxiv.2412.15687

  73. [78]

    and Reynolds, Carolyn A

    Baño-Medina, Jorge and Sengupta, Agniv and Doyle, James D. and Reynolds, Carolyn A. and Watson-Parris, Duncan and Monache, Luca Delle , month = mar, year =. Are. npj Climate and Atmospheric Science , publisher =. doi:10.1038/s41612-025-00949-6 , language =

    work page doi:10.1038/s41612-025-00949-6

  74. [79]

    , month = aug, year =

    Nichani, Eshaan and Damian, Alex and Lee, Jason D. , month = aug, year =. How. doi:10.48550/arXiv.2402.14735 , abstract =

    work page doi:10.48550/arxiv.2402.14735

  75. [80]

    Lang, Simon and Alexe, Mihai and Clare, Mariana C. A. and Roberts, Christopher and Adewoyin, Rilwan and Ben Bouallègue, Zied and Chantry, Matthew and Dramsch, Jesper and Dueben, Peter D. and Hahner, Sara and Maciel, Pedro and Prieto-Nemesio, Ana and O’Brien, Cathal and Pinault, Florian and Polster, Jan and Raoult, Baudouin and Tietsche, Steffen and Leutbe...

    work page doi:10.1038/s44387-026-00073-7

  76. [81]

    Fixing the

    Subich, Christopher and Husain, Syed Zahid and Separovic, Leo and Yang, Jing , month = jan, year =. Fixing the. doi:10.48550/arXiv.2501.19374 , abstract =

    work page doi:10.48550/arxiv.2501.19374

  77. [82]

    Computer Methods in Applied Mechanics and Engineering , author =

    Can diffusion models capture extreme event statistics? , volume =. Computer Methods in Applied Mechanics and Engineering , author =. 2025 , keywords =. doi:10.1016/j.cma.2024.117589 , abstract =

    work page doi:10.1016/j.cma.2024.117589 2025

  78. [83]

    URL https://www.nature.com/articles/ s41586-024-07744-y

    Kochkov, Dmitrii and Yuval, Janni and Langmore, Ian and Norgaard, Peter and Smith, Jamie and Mooers, Griffin and Klöwer, Milan and Lottes, James and Rasp, Stephan and Düben, Peter and Hatfield, Sam and Battaglia, Peter and Sanchez-Gonzalez, Alvaro and Willson, Matthew and Brenner, Michael P. and Hoyer, Stephan , month = aug, year =. Neural general circula...

    work page doi:10.1038/s41586-024-07744-y

  79. [84]

    Geophysical Research Letters , author =

    Explainable. Geophysical Research Letters , author =. 2024 , note =. doi:10.1029/2023GL106324 , abstract =

    work page doi:10.1029/2023gl106324 2024

  80. [85]

    Eulalie Boucher, Mihai Alexe, Peter Lean, Ewan Pinnington, Simon Lang, Patrick Laloyaux, Lorenzo Zampieri, Patricia de Rosnay, Niels Bormann, and Anthony McNally

    Bi, Kaifeng and Xie, Lingxi and Zhang, Hengheng and Chen, Xin and Gu, Xiaotao and Tian, Qi , month = jul, year =. Accurate medium-range global weather forecasting with. Nature , publisher =. doi:10.1038/s41586-023-06185-3 , language =

    work page doi:10.1038/s41586-023-06185-3

Showing first 80 references.