pith. sign in

arxiv: 2605.31204 · v1 · pith:YXHX3UFQnew · submitted 2026-05-29 · 💻 cs.CV

Probabilistic Precipitation Nowcasting with Rectified Flow Transformers

Johannes Schusterbauer , Jannik Wiese , Nick Stracke , Timy Phan , Bj\"orn Ommer This is my paper

Pith reviewed 2026-06-28 23:13 UTC · model grok-4.3

classification 💻 cs.CV
keywords precipitation nowcastingrectified flow transformersprobabilistic forecastingweather predictionSEVIRdiffusion modelsspatio-temporal dataaleatoric uncertainty
0
0 comments X

The pith

A frame-wise encoder and united decoder with rectified flow transformers enables state-of-the-art probabilistic precipitation nowcasting by preserving uncertainty.

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

The paper presents FREUD, which uses a frame-wise encoder to compress spatio-temporal weather data and a unified decoder based on rectified flow transformers. This design supports continuous updates and ensures temporal consistency in forecasts. By keeping uncertainty in the first stage, ensembling can capture aleatoric uncertainty, improving predictions especially for extreme weather. The approach reaches state-of-the-art on the SEVIR benchmark and benefits from model and test-time scaling.

Core claim

FREUD achieves state-of-the-art performance in precipitation nowcasting with a compact latent-space rectified flow transformer on the SEVIR benchmark by using an uncertainty-preserving first stage that allows capture of aleatoric uncertainty via ensembling.

What carries the argument

FREUD, the Frame-wise Encoder and United Decoder model based on rectified flow transformers, which enables efficient compression while preserving uncertainty for ensembling.

If this is right

  • Frame-wise encoding permits continuous forecast updates without full reprocessing.
  • The unified video decoder maintains temporal consistency across the forecast sequence.
  • Ensembling from the uncertainty-preserving stage particularly improves forecasts for extreme weather events.
  • Model scaling and test-time scaling yield additional performance gains.

Where Pith is reading between the lines

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

  • Similar uncertainty-preserving compression could apply to other high-dimensional spatio-temporal data tasks like video prediction.
  • Test-time scaling might be combined with other probabilistic methods for further gains in forecasting.
  • The method's efficiency could enable real-time nowcasting in operational weather systems.

Load-bearing premise

That the uncertainty in the latent space from the first stage can be effectively leveraged through ensembling to improve accuracy on extreme events.

What would settle it

Demonstrating that ensembling decodings from the uncertainty-preserving stage does not improve performance on extreme weather events compared to deterministic methods would falsify the key benefit.

Figures

Figures reproduced from arXiv: 2605.31204 by Bj\"orn Ommer, Jannik Wiese, Johannes Schusterbauer, Nick Stracke, Timy Phan.

Figure 1
Figure 1. Figure 1: Reconstruction ensemble distributions from our gen￾erative compression stage for locations with medium (blue), heavy (orange), and extreme (red) precipitation. In light-rain regions, the ensemble shows low variance, while under intense and chaotic rainfall, the spread increases and reliably covers the ground-truth values (dots), whereas the deterministic CasCast decoder [40] (lines) cannot quantify uncerta… view at source ↗
Figure 2
Figure 2. Figure 2: FREUD architecture. Left: Frame-wise encoder. Right: Generative decoder conditioned on encoder latents z. 153], often biases the generation [98]. We address these issues with a first stage that quantifies compression uncer￾tainty and enables purely data-driven nowcasting. Further discussion of related works is provided in the Appendix. Uncertainty Estimation in ML differentiates uncertainty from underdeter… view at source ↗
Figure 3
Figure 3. Figure 3: Inference pipeline. We first generate weather forecasts in the latent space of our FREUD encoder and then decode the predictions with our generative FREUD decoder. an error because of the stochastic component. Therefore, if we can draw Monte-Carlo samples x out e from the distribu￾tion, the sample variance \text {Var}(\mathbf {\tilde {x}}^{\text {out}}),\quad \mathbf {\tilde x}^{\text {out}} = [\mathbf {x}… view at source ↗
Figure 4
Figure 4. Figure 4: Forecast ensemble from our L model. Red rectangles show zoom-ins to highlight differences. Best viewed zoomed in. decoder jointly with the RF loss from Eq. (2). Latent Space Regularization A well-structured latent space is essential for latent space generative modeling [43, 152, 161]. Current autoencoders typically include a small KL regularization term to encourage a smooth latent space [110]. In contrast… view at source ↗
Figure 5
Figure 5. Figure 5: FREUD efficiency. Our transformer-based autoencoder has fewer parameters and uses fewer FLOPs for encoding and de￾coding (5 NFE). This allows for faster training and inference. the FREUD first stage with T-reg. is thus just the flow loss from Eq. (2), simplified with the linear schedule to \mathcal {L} = \lVert \mathbf {v}_\theta (\mathbf {x}_i, i) - (\mathbf {x}_1 - \mathbf {x}_0) \rVert ^2 (5) where x1 i… view at source ↗
Figure 6
Figure 6. Figure 6: Correlation between precipitation intensity and de￾coding variance. All FREUD variants show a strong linear rela￾tionship, with T-reg. achieving the highest correlation. reg. FREUD achieves superior results across all metrics, with a significant gain in SSIM, indicating perceptually accurate reconstructions without adversarial or perceptual losses. It further needs fewer parameters and fewer Floating Point… view at source ↗
Figure 7
Figure 7. Figure 7: Qualitative reconstructions of a normal and an ex￾treme weather event, with corresponding ensemble members and ensemble variance. The red rectangle highlights a zoomed-in area for visualizing differences in details. Best viewed zoomed in. in high-impact regions. The variance maps reveal a clear correlation between precipitation intensity and reconstruc￾tion variance, consistent with the non-linearity and c… view at source ↗
Figure 8
Figure 8. Figure 8: Calibration Rank Histogram for our pipeline. Our rank histogram is flatter, indicating improved calibration com￾pared to CasCast [40] [PITH_FULL_IMAGE:figures/full_fig_p008_8.png] view at source ↗
read the original abstract

Accurate weather forecasts are essential across various domains and are safety-critical in extreme weather conditions. Compared to simulation-based forecasting, data-driven approaches show greater efficiency, enabling short-term, high-resolution nowcasting. In particular, diffusion models proved effective in weather nowcasting due to their strong probabilistic foundation. However, existing methods rely on deterministic compression to reduce the complexity of high-dimensional weather data, limiting their ability to capture uncertainty in the decoding process. In this work, we introduce $\textbf{FREUD}$, a $\textbf{Fr}$ame-wise $\textbf{E}$ncoder and $\textbf{U}$nited $\textbf{D}$ecoder model based on rectified flow transformers for efficient compression of spatio-temporal weather data. Frame-wise encoding enables continuous forecast updates, while the unified video decoder ensures temporal consistency. Our uncertainty-preserving first stage allows us to capture aleatoric uncertainty via ensembling, which is particularly beneficial for extreme weather events with high decoding variability. We achieve state-of-the-art performance in precipitation nowcasting with a compact latent-space rectified flow transformer on the SEVIR benchmark and show further performance gains by model and test-time scaling. Code available here: https://github.com/CompVis/weather-rf

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

0 major / 1 minor

Summary. The paper claims to introduce FREUD, a frame-wise encoder and united decoder model based on rectified flow transformers for efficient compression of spatio-temporal weather data in precipitation nowcasting. It highlights the limitation of deterministic compression in existing methods and proposes an uncertainty-preserving first stage to capture aleatoric uncertainty via ensembling, achieving state-of-the-art performance on the SEVIR benchmark with additional gains from scaling.

Significance. If the empirical claims hold, this work is significant for the field of data-driven weather forecasting as it offers a compact latent-space model with temporal consistency and uncertainty awareness, potentially improving forecasts for extreme weather events. The release of code is a notable strength for reproducibility and further research. The stress-test concern on the aleatoric uncertainty claim via ensembling does not land as a load-bearing flaw.

minor comments (1)
  1. The GitHub link is provided; ensure it remains accessible and consider adding a DOI or archive link for long-term availability.

Simulated Author's Rebuttal

0 responses · 0 unresolved

We thank the referee for their thorough review and positive recommendation to accept the manuscript. We appreciate the recognition of FREUD's contributions to probabilistic precipitation nowcasting, particularly the uncertainty-preserving compression approach and the benefits of code release for reproducibility.

Circularity Check

0 steps flagged

No significant circularity

full rationale

The paper's central claims consist of empirical benchmark results (SOTA on SEVIR via a latent-space rectified flow transformer, plus scaling gains) and an architectural description of FREUD. No derivation chain, uniqueness theorem, or first-principles prediction is presented that reduces by construction to fitted inputs, self-citations, or renamed known results. The uncertainty-preserving encoder claim is an interpretive modeling choice, not a load-bearing reduction. All performance assertions rest on external benchmarks and code, making the argument self-contained against verifiable experiments.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Abstract-only; no explicit free parameters, axioms, or invented entities are stated or derivable.

pith-pipeline@v0.9.1-grok · 5751 in / 997 out tokens · 20454 ms · 2026-06-28T23:13:16.486689+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

164 extracted references · 11 canonical work pages · 3 internal anchors

  1. [1]

    Machine learn- ing for precipitation nowcasting from radar images.arXiv preprint arXiv:1912.12132, 2019

    Shreya Agrawal, Luke Barrington, Carla Bromberg, John Burge, Cenk Gazen, and Jason Hickey. Machine learn- ing for precipitation nowcasting from radar images.arXiv preprint arXiv:1912.12132, 2019. 1, 2, 10, 11

    work page arXiv 1912

  2. [2]

    Al-Saegh, Taha A

    Ali M. Al-Saegh, Taha A. Elwi, Osama A. Abdullah, Aduwati Sali, and Abdulmajeed H. J. Aljumaily. Rainfall Effect on Satellite Communications in Mosul at Frequen- cies above 10 GHz. In2021 7th International Conference on Space Science and Communication (IconSpace), pages 318–322, 2021. 2

    2021

  3. [3]

    Build- ing normalizing flows with stochastic interpolants

    Michael Samuel Albergo and Eric Vanden-Eijnden. Build- ing normalizing flows with stochastic interpolants. InICLR,

  4. [4]

    Stochastic Interpolants: A Unifying Framework for Flows and Diffusions

    Michael S Albergo, Nicholas M Boffi, and Eric Vanden- Eijnden. Stochastic interpolants: A unifying framework for flows and diffusions.arXiv preprint arXiv:2303.08797,

    work page internal anchor Pith review Pith/arXiv arXiv

  5. [5]

    Andersson, Jacklynn Stott, Remi Lam, Matthew Willson, Alvaro Sanchez-Gonzalez, and Pe- ter Battaglia

    Ferran Alet, Ilan Price, Andrew El-Kadi, Dominic Masters, Stratis Markou, Tom R. Andersson, Jacklynn Stott, Remi Lam, Matthew Willson, Alvaro Sanchez-Gonzalez, and Pe- ter Battaglia. Skillful joint probabilistic weather forecasting from marginals, 2025. 11

    2025

  6. [6]

    Latent-Shift: Latent Diffusion with Temporal Shift for Efficient Text-to-Video Generation, 2023

    Jie An, Songyang Zhang, Harry Yang, Sonal Gupta, Jia- Bin Huang, Jiebo Luo, and Xi Yin. Latent-Shift: Latent Diffusion with Temporal Shift for Efficient Text-to-Video Generation, 2023. 3

    2023

  7. [7]

    Precipitation nowcasting with generative diffusion models,

    Andrea Asperti, Fabio Merizzi, Alberto Paparella, Gior- gio Pedrazzi, Matteo Angelinelli, and Stefano Colamonaco. Precipitation nowcasting with generative diffusion models,

  8. [8]

    4M-21: An Any-to-Any Vision Model for Tens of Tasks and Modalities

    Roman Bachmann, O ˘guzhan Fatih Kar, David Mizrahi, Ali Garjani, Mingfei Gao, David Griffiths, Jiaming Hu, Afshin Dehghan, and Amir Zamir. 4M-21: An Any-to-Any Vision Model for Tens of Tasks and Modalities. InThe Thirty- eighth Annual Conference on Neural Information Process- ing Systems, 2024. 3

    2024

  9. [9]

    FlexTok: Resampling Images into 1D Token Sequences of Flexible Length, 2025

    Roman Bachmann, Jesse Allardice, David Mizrahi, En- rico Fini, O ˘guzhan Fatih Kar, Elmira Amirloo, Alaaeldin El-Nouby, Amir Zamir, and Afshin Dehghan. FlexTok: Resampling Images into 1D Token Sequences of Flexible Length, 2025. 3

    2025

  10. [10]

    Is Space-Time Attention All You Need for Video Understand- ing?, 2021

    Gedas Bertasius, Heng Wang, and Lorenzo Torresani. Is Space-Time Attention All You Need for Video Understand- ing?, 2021. 7

    2021

  11. [11]

    Improving Image Generation with Better Captions

    James Betker, Gabriel Goh, Li Jing, † TimBrooks, Jian- feng Wang, Linjie Li, † LongOuyang, † JuntangZhuang, † JoyceLee, † YufeiGuo, † WesamManassra, † PrafullaD- hariwal, † CaseyChu, † YunxinJiao, and Aditya Ramesh. Improving Image Generation with Better Captions. InCom- puter Science, 2023. 3

    2023

  12. [12]

    Align Your Latents: High-Resolution Video Syn- thesis with Latent Diffusion Models

    Andreas Blattmann, Robin Rombach, Huan Ling, Tim Dockhorn, Seung Wook Kim, Sanja Fidler, and Karsten Kreis. Align Your Latents: High-Resolution Video Syn- thesis with Latent Diffusion Models. In2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 22563–22575, Vancouver, BC, Canada,

  13. [13]

    Towards replacing precip- itation ensemble predictions systems using machine learn- ing.Atmospheric Science Letters, n/a(n/a):e1262, 2024

    R ¨udiger Brecht and Alex Bihlo. Towards replacing precip- itation ensemble predictions systems using machine learn- ing.Atmospheric Science Letters, n/a(n/a):e1262, 2024. 7

    2024

  14. [14]

    Chan, Maria J

    Matthew A. Chan, Maria J. Molina, and Christopher A. Metzler. Hyper-Diffusion: Estimating Epistemic and Aleatoric Uncertainty with a Single Model, 2024. 1, 3, 4, 7

    2024

  15. [15]

    Diffusion forcing: Next-token prediction meets full-sequence diffu- sion, 2024

    Boyuan Chen, Diego Marti Monso, Yilun Du, Max Sim- chowitz, Russ Tedrake, and Vincent Sitzmann. Diffusion forcing: Next-token prediction meets full-sequence diffu- sion, 2024. 6

    2024

  16. [16]

    Diffusion Forcing: Next-token prediction meets full-sequence diffu- sion, 2024

    Boyuan Chen, Diego Marti Monso, Yilun Du, Max Sim- chowitz, Russ Tedrake, and Vincent Sitzmann. Diffusion Forcing: Next-token prediction meets full-sequence diffu- sion, 2024. 2

    2024

  17. [17]

    Lungren, Shaoting Zhang, Lei Xing, Le Lu, Alan Yuille, and Yuyin Zhou

    Jieneng Chen, Jieru Mei, Xianhang Li, Yongyi Lu, Qihang Yu, Qingyue Wei, Xiangde Luo, Yutong Xie, Ehsan Adeli, Yan Wang, Matthew P. Lungren, Shaoting Zhang, Lei Xing, Le Lu, Alan Yuille, and Yuyin Zhou. TransUNet: Rethink- ing the U-Net architecture design for medical image seg- mentation through the lens of transformers.Medical Image Analysis, 97:103280,...

    2024

  18. [18]

    Lei Chen, Fei Du, Yuan Hu, Zhibin Wang, and Fan Wang. SwinRDM: Integrate SwinRNN with Diffusion Model to- wards High-Resolution and High-Quality Weather Fore- casting.Proceedings of the AAAI Conference on Artificial Intelligence, 37(1):322–330, 2023. 12

    2023

  19. [19]

    Neural ordinary differential equa- tions.Advances in neural information processing systems, 31, 2018

    Ricky TQ Chen, Yulia Rubanova, Jesse Bettencourt, and David K Duvenaud. Neural ordinary differential equa- tions.Advances in neural information processing systems, 31, 2018. 3, 11

    2018

  20. [20]

    Foundation Models for Weather and Climate Data Understanding: A Comprehensive Survey,

    Shengchao Chen, Guodong Long, Jing Jiang, Dikai Liu, and Chengqi Zhang. Foundation Models for Weather and Climate Data Understanding: A Comprehensive Survey,

  21. [21]

    Diffusion Autoencoders are Scalable Image Tokenizers, 2025

    Yinbo Chen, Rohit Girdhar, Xiaolong Wang, Sai Saketh Rambhatla, and Ishan Misra. Diffusion Autoencoders are Scalable Image Tokenizers, 2025. 3, 4, 6

    2025

  22. [22]

    Cheung and H.Y

    P. Cheung and H.Y . Yeung. Application of optical-flow technique to significant convection nowcast for terminal ar- eas in Hong Kong. InThe 3rd WMO International Sympo- sium on Nowcasting and Very Short-Range Forecastin, Rio de Janeiro, Brazil, 2012. WMO. 2, 10

    2012

  23. [23]

    Pembleton

    Brian Collins, Yunru Lai, Uwe Grewer, Steve Attard, Justin Sexton, and Keith G. Pembleton. Evaluating the impact of weather forecasts on productivity and environmental foot- print of irrigated maize production systems.Science of The Total Environment, 954:176368, 2024. 1

    2024

  24. [24]

    Kaplan, and Enrico Shippole

    Katherine Crowson, Stefan Andreas Baumann, Alex Birch, Tanishq Mathew Abraham, Daniel Z. Kaplan, and Enrico Shippole. Scalable High-Resolution Pixel-Space Image Synthesis with Hourglass Diffusion Transformers. InPro- ceedings of the 41st International Conference on Machine Learning, pages 9550–9575. PMLR, 2024. 2, 4, 7

    2024

  25. [25]

    Saurabh Das, Animesh Maitra, and Ashish K. Shukla. Rain Attenuation Modeling in the 10-100 GHz Frequency us- ing Drop Size Distributions for different Climatic Zones in Tropical India.Progress In Electromagnetics Research B, 25:211–224, 2010. 2

    2010

  26. [26]

    Hacker, Yongmei Zhou, Xingxiu Deng, and Roland B

    Luca Delle Monache, Joshua P. Hacker, Yongmei Zhou, Xingxiu Deng, and Roland B. Stull. Probabilistic aspects of meteorological and ozone regional ensemble forecasts. Journal of Geophysical Research: Atmospheres, 111(D24): 2005JD006917, 2006. 7

    2006

  27. [27]

    Density estimation using real nvp

    Laurent Dinh, Jascha Sohl-Dickstein, and Samy Bengio. Density estimation using real nvp. InInternational Con- ference on Learning Representations, 2017. 3

    2017

  28. [28]

    An image is worth 16x16 words: Transformers for image recognition at scale

    Alexey Dosovitskiy, Lucas Beyer, Alexander Kolesnikov, Dirk Weissenborn, Xiaohua Zhai, Thomas Unterthiner, Mostafa Dehghani, Matthias Minderer, G Heigold, S Gelly, et al. An image is worth 16x16 words: Transformers for image recognition at scale. InICLR, 2020. 7, 10

    2020

  29. [29]

    Dowell, Curtis R

    David C. Dowell, Curtis R. Alexander, Eric P. James, Stephen S. Weygandt, Stanley G. Benjamin, Geoffrey S. Manikin, Benjamin T. Blake, John M. Brown, Joseph B. Olson, Ming Hu, Tatiana G. Smirnova, Terra Ladwig, Jaymes S. Kenyon, Ravan Ahmadov, David D. Turner, Jef- frey D. Duda, and Trevor I. Alcott. The High-Resolution Rapid Refresh (HRRR): An Hourly Upd...

  30. [30]

    Analysis of weather-driven air traffic management challenges for major us and euro- pean airports

    Gabriele Enea, Tom Reynolds, Mark Weber, Ramon Dal- mau Codina, and Dirk Schaefer. Analysis of weather-driven air traffic management challenges for major us and euro- pean airports. InSESAR Innovation Days, Rome, 2024. SESAR Joint Undertaking. 1

    2024

  31. [31]

    Deep learning for twelve hour precipitation forecasts.Nature communications, 13(1):5145, 2022

    Lasse Espeholt, Shreya Agrawal, Casper Sønderby, Manoj Kumar, Jonathan Heek, Carla Bromberg, Cenk Gazen, Rob Carver, Marcin Andrychowicz, Jason Hickey, et al. Deep learning for twelve hour precipitation forecasts.Nature communications, 13(1):5145, 2022. 1, 10

    2022

  32. [32]

    Scaling rectified flow transformers for high-resolution image synthesis

    Patrick Esser, Sumith Kulal, Andreas Blattmann, Rahim Entezari, Jonas M ¨uller, Harry Saini, Yam Levi, Dominik Lorenz, Axel Sauer, Frederic Boesel, et al. Scaling rectified flow transformers for high-resolution image synthesis. In ICML, 2024. 3

    2024

  33. [33]

    Broad- unet: Multi-scale feature learning for nowcasting tasks

    Jes ´us Garc´ıa Fern´andez and Siamak Mehrkanoon. Broad- unet: Multi-scale feature learning for nowcasting tasks. Neural Networks, 144:419–427, 2021. 1, 10

    2021

  34. [34]

    Dropout as a Bayesian Approximation: Representing Model Uncertainty in Deep Learning

    Yarin Gal and Zoubin Ghahramani. Dropout as a Bayesian Approximation: Representing Model Uncertainty in Deep Learning, 2016. arXiv:1506.02142. 3

    work page internal anchor Pith review Pith/arXiv arXiv 2016

  35. [35]

    Earth- former: Exploring space-time transformers for earth system forecasting.Advances in Neural Information Processing Systems, 35:25390–25403, 2022

    Zhihan Gao, Xingjian Shi, Hao Wang, Yi Zhu, Yuyang Bernie Wang, Mu Li, and Dit-Yan Yeung. Earth- former: Exploring space-time transformers for earth system forecasting.Advances in Neural Information Processing Systems, 35:25390–25403, 2022. 2, 3, 6, 7, 8, 10, 11, 12, 21, 22

    2022

  36. [36]

    Simvp: Simpler yet better video prediction

    Zhangyang Gao, Cheng Tan, Lirong Wu, and Stan Z Li. Simvp: Simpler yet better video prediction. InProceed- ings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 3170–3180, 2022. 2, 7, 10, 12

    2022

  37. [37]

    Prediff: Precipitation nowcasting with latent diffusion models

    Zhihan Gao, Xingjian Shi, Boran Han, Hao Wang, Xi- aoyong Jin, Danielle Maddix, Yi Zhu, Mu Li, and Yuyang (Bernie) Wang. Prediff: Precipitation nowcasting with latent diffusion models. InAdvances in Neural In- formation Processing Systems, pages 78621–78656. Curran Associates, Inc., 2023. 1, 2, 3, 7, 10, 11, 12

    2023

  38. [38]

    Preserve Your Own Correlation: A Noise Prior for Video Diffusion Models

    Songwei Ge, Seungjun Nah, Guilin Liu, Tyler Poon, An- drew Tao, Bryan Catanzaro, David Jacobs, Jia-Bin Huang, Ming-Yu Liu, and Yogesh Balaji. Preserve Your Own Correlation: A Noise Prior for Video Diffusion Models. In2023 IEEE/CVF International Conference on Computer Vision (ICCV), pages 22873–22884, Paris, France, 2023. IEEE. 3

    2023

  39. [39]

    Scale-Dependence of the Predictability of Precipitation from Continental Radar Images

    Urs Germann and Isztar Zawadzki. Scale-Dependence of the Predictability of Precipitation from Continental Radar Images. Part I: Description of the Methodology.Monthly Weather Review, 130(12):2859–2873, 2002. 2, 10

    2002

  40. [40]

    CasCast: Skillful high-resolution precipitation nowcasting via cas- caded modelling

    Junchao Gong, Lei Bai, Peng Ye, Wanghan Xu, Na Liu, Jianhua Dai, Xiaokang Yang, and Wanli Ouyang. CasCast: Skillful high-resolution precipitation nowcasting via cas- caded modelling. InProceedings of the 41st International Conference on Machine Learning, pages 15809–15822, Vi- enna, Austria, 2024. JMLR.org. 1, 2, 4, 6, 7, 8, 3, 5, 10, 12, 16, 17, 21, 22

    2024

  41. [41]

    Generative adversarial nets

    Ian Goodfellow, Jean Pouget-Abadie, Mehdi Mirza, Bing Xu, David Warde-Farley, Sherjil Ozair, Aaron Courville, and Yoshua Bengio. Generative adversarial nets. InAd- vances in Neural Information Processing Systems. Curran Associates, Inc., 2014. 1, 2, 11

    2014

  42. [42]

    Disentangling physi- cal dynamics from unknown factors for unsupervised video prediction

    Vincent Le Guen and Nicolas Thome. Disentangling physi- cal dynamics from unknown factors for unsupervised video prediction. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 11474– 11484, 2020. 2, 7, 10, 11, 12

    2020

  43. [43]

    Adapting self-supervised representations as a la- tent space for efficient generation, 2025

    Ming Gui, Johannes Schusterbauer, Timy Phan, Felix Krause, Josh Susskind, Miguel Angel Bautista, and Bj ¨orn Ommer. Adapting self-supervised representations as a la- tent space for efficient generation, 2025. 2, 4, 5

    2025

  44. [44]

    Flexible diffusion model- ing of long videos.arXiv preprint arXiv:2205.11495, 2022

    William Harvey, Saeid Naderiparizi, Vaden Masrani, Chris- tian Weilbach, and Frank Wood. Flexible diffusion model- ing of long videos.arXiv preprint arXiv:2205.11495, 2022. 4

    work page arXiv 2022

  45. [45]

    Neighborhood Attention Transformer

    Ali Hassani, Steven Walton, Jiachen Li, Shen Li, and Humphrey Shi. Neighborhood Attention Transformer. In Proceedings of the IEEE/CVF Conference on Computer Vi- sion and Pattern Recognition, pages 6185–6194, 2023. 4, 7

    2023

  46. [46]

    Latent Video Diffusion Models for High- Fidelity Long Video Generation, 2023

    Yingqing He, Tianyu Yang, Yong Zhang, Ying Shan, and Qifeng Chen. Latent Video Diffusion Models for High- Fidelity Long Video Generation, 2023. 3

    2023

  47. [47]

    Rotary Position Embedding for Vision Transformer,

    Byeongho Heo, Song Park, Dongyoon Han, and Sangdoo Yun. Rotary Position Embedding for Vision Transformer,

  48. [48]

    Classifier-Free Diffusion Guidance, 2022

    Jonathan Ho and Tim Salimans. Classifier-Free Diffusion Guidance, 2022. 6, 7, 3, 4, 5, 9, 11, 14, 15, 16, 17

    2022

  49. [49]

    Axial Attention in Multidimensional Trans- formers, 2019

    Jonathan Ho, Nal Kalchbrenner, Dirk Weissenborn, and Tim Salimans. Axial Attention in Multidimensional Trans- formers, 2019. 10

    2019

  50. [50]

    Denoising dif- fusion probabilistic models.Advances in Neural Informa- tion Processing Systems, 33:6840–6851, 2020

    Jonathan Ho, Ajay Jain, and Pieter Abbeel. Denoising dif- fusion probabilistic models.Advances in Neural Informa- tion Processing Systems, 33:6840–6851, 2020. 1, 3, 11

    2020

  51. [51]

    Long Short- Term Memory.Neural Comput., 9(8):1735–1780, 1997

    Sepp Hochreiter and J ¨urgen Schmidhuber. Long Short- Term Memory.Neural Comput., 9(8):1735–1780, 1997. 10

    1997

  52. [52]

    Yang Hong, Kuo-Lin Hsu, Soroosh Sorooshian, and Xiao- gang Gao. Precipitation Estimation from Remotely Sensed Imagery Using an Artificial Neural Network Cloud Classi- fication System.Journal of Applied Meteorology and Cli- matology, 43(12):1834–1853, 2004. 9

    2004

  53. [53]

    Yang Hong, Kuo-lin Hsu, Hamid Moradkhani, and Soroosh Sorooshian. Uncertainty quantification of satellite precipi- tation estimation and Monte Carlo assessment of the error propagation into hydrologic response.Water Resources Re- search, 42(8), 2006. 3

    2006

  54. [54]

    Diffusion models for video prediction and infilling.arXiv preprint arXiv:2206.07696,

    Tobias H ¨oppe, Arash Mehrjou, Stefan Bauer, Didrik Nielsen, and Andrea Dittadi. Diffusion models for video prediction and infilling.arXiv preprint arXiv:2206.07696,

    work page arXiv

  55. [55]

    MiniCPM: Unveiling the Potential of Small Language Models with Scalable Training Strategies, 2024

    Shengding Hu, Yuge Tu, Xu Han, Chaoqun He, Ganqu Cui, Xiang Long, Zhi Zheng, Yewei Fang, Yuxiang Huang, Weilin Zhao, Xinrong Zhang, Zheng Leng Thai, Kai- huo Zhang, Chongyi Wang, Yuan Yao, Chenyang Zhao, Jie Zhou, Jie Cai, Zhongwu Zhai, Ning Ding, Chao Jia, Guoyang Zeng, Dahai Li, Zhiyuan Liu, and Maosong Sun. MiniCPM: Unveiling the Potential of Small Lan...

    2024

  56. [56]

    Hudson, Daniel Zoran, Mateusz Malinowski, An- drew K

    Drew A. Hudson, Daniel Zoran, Mateusz Malinowski, An- drew K. Lampinen, Andrew Jaegle, James L. McClelland, Loic Matthey, Felix Hill, and Alexander Lerchner. SODA: Bottleneck Diffusion Models for Representation Learning. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 23115–23127, 2024. 3

    2024

  57. [57]

    Aleatoric and epistemic uncertainty in machine learning: An introduction to concepts and methods.Machine Learning, 110(3):457– 506, 2021

    Eyke H ¨ullermeier and Willem Waegeman. Aleatoric and epistemic uncertainty in machine learning: An introduction to concepts and methods.Machine Learning, 110(3):457– 506, 2021. 3, 11

    2021

  58. [58]

    Iterative Residual Refinement for Joint Optical Flow and Occlusion Estimation, 2019

    Junhwa Hur and Stefan Roth. Iterative Residual Refinement for Joint Optical Flow and Occlusion Estimation, 2019. 2

    2019

  59. [59]

    Imhoff, Lesley De Cruz, Wout Dewettinck, Clau- dia C

    Ruben O. Imhoff, Lesley De Cruz, Wout Dewettinck, Clau- dia C. Brauer, Remko Uijlenhoet, Klaas-Jan Van Heerin- gen, Carlos Velasco-Forero, Daniele Nerini, Michiel Van Ginderachter, and Albrecht H. Weerts. Scale- dependent blending of ensemble rainfall nowcasts and nu- merical weather prediction in the open-source pysteps li- brary.Quarterly Journal of the...

    2023

  60. [60]

    Users’ guide to imagery with heavy rainfall potential areas, 2014

    Japan Meteorological Agency. Users’ guide to imagery with heavy rainfall potential areas, 2014. 9

    2014

  61. [61]

    Clgan: A gan-based video prediction model for precipitation nowcasting.EGUsphere, pages 1– 23, 2022

    Yan Ji, Bing Gong, Michael Langguth, Amirpasha Mozaf- fari, and Xiefei Zhi. Clgan: A gan-based video prediction model for precipitation nowcasting.EGUsphere, pages 1– 23, 2022. 11

    2022

  62. [62]

    TC-Diffusion: A diffusion-based probabilistic tropical cyclone model with application to typhoon wind hazard assessment.Reliability Engineering & System Safety, 251:110350, 2024

    Wenjun Jiang, Xi Zhong, and Jize Zhang. TC-Diffusion: A diffusion-based probabilistic tropical cyclone model with application to typhoon wind hazard assessment.Reliability Engineering & System Safety, 251:110350, 2024. 2

    2024

  63. [63]

    Kanellopoulos, Stelios G

    John D. Kanellopoulos, Stelios G. Koukoulas, Nicolaos J. Kolliopoulos, Christos N. Capsalis, and Spyros G. Ven- touras. Rain attenuation problems affecting the perfor- mance of microwave communication systems.Annales des T´el´ecommunications, 45(7):437–451, 1990. 2

    1990

  64. [64]

    Khalil.Nonlinear Systems

    Hassan K. Khalil.Nonlinear Systems. Prentice Hall, Upper Saddle River, NJ, 3. ed edition, 2002. 10

    2002

  65. [65]

    Kingma and Max Welling

    Diederik P. Kingma and Max Welling. Auto-encoding vari- ational bayes. In2nd International Conference on Learning Representations, ICLR 2014, Banff, AB, Canada, April 14- 16, 2014, Conference Track Proceedings, 2014. 2

    2014

  66. [66]

    Mode Collapse in Generative Adversarial Networks: An Overview

    Youssef Kossale, Mohammed Airaj, and Aziz Darouichi. Mode Collapse in Generative Adversarial Networks: An Overview. In2022 8th International Conference on Op- timization and Applications (ICOA), pages 1–6, 2022. 1, 2, 4, 11

    2022

  67. [67]

    Simple and scalable predictive uncertainty esti- mation using deep ensembles

    Balaji Lakshminarayanan, Alexander Pritzel, and Charles Blundell. Simple and scalable predictive uncertainty esti- mation using deep ensembles. InProceedings of the 31st International Conference on Neural Information Process- ing Systems, pages 6405–6416, Red Hook, NY , USA, 2017. Curran Associates Inc. 3

    2017

  68. [68]

    Learning skillful medium- range global weather forecasting.Science, 2023

    Remi Lam, Alvaro Sanchez-Gonzalez, Matthew Willson, Peter Wirnsberger, Meire Fortunato, Ferran Alet, Suman Ravuri, Timo Ewalds, Zach Eaton-Rosen, Weihua Hu, Alexander Merose, Stephan Hoyer, George Holland, Oriol Vinyals, Jacklynn Stott, Alexander Pritzel, Shakir Mo- hamed, and Peter Battaglia. Learning skillful medium- range global weather forecasting.Sci...

    2023

  69. [69]

    Fluid Mechanics

    Lev Davidovich Landau and Evgeni˘ı Mikha˘ılovich Lifshits. Fluid Mechanics. Number v. 6 in Their Course of Theoret- ical Physics. Pergamon Press, London, 1959. 10

    1959

  70. [70]

    Meteonet, an open reference weather dataset by meteo-france.URL https://meteonet.umr-cnrm.fr/, 2020

    Gwenna ¨elle Larvor, L´ea Berthomier, Vincent Chabot, Brice Le Pape, Bruno Pradel, and Lior Perez. Meteonet, an open reference weather dataset by meteo-france.URL https://meteonet.umr-cnrm.fr/, 2020. 6, 3

    2020

  71. [71]

    Leinonen, U

    Jussi Leinonen, Ulrich Hamann, Daniele Nerini, Urs Ger- mann, and Gabriele Franch. Latent diffusion models for generative precipitation nowcasting with accurate uncer- tainty quantification.arXiv preprint arXiv:2304.12891,

    work page arXiv

  72. [72]

    1, 2, 3, 4, 7, 10, 11, 12

  73. [73]

    SEEDS: Emulation of Weather Forecast Ensembles with Diffusion Models, 2023

    Lizao Li, Rob Carver, Ignacio Lopez-Gomez, Fei Sha, and John Anderson. SEEDS: Emulation of Weather Forecast Ensembles with Diffusion Models, 2023. 12

    2023

  74. [74]

    Generative emulation of weather fore- cast ensembles with diffusion models.Science Advances, 10(13):eadk4489, 2024

    Lizao Li, Robert Carver, Ignacio Lopez-Gomez, Fei Sha, and John Anderson. Generative emulation of weather fore- cast ensembles with diffusion models.Science Advances, 10(13):eadk4489, 2024. 7

    2024

  75. [75]

    Latent Diffusion Model for Quantitative Precipitation Estimation and Fore- cast at km Scale.ESS Open Archive, 2024

    Weidong Li, Baoxiang Pan, Tiejian Li, Congyi Nai, Zhaoxi Li, Bo Lu, Qingyun Duan, and Ming Pan. Latent Diffusion Model for Quantitative Precipitation Estimation and Fore- cast at km Scale.ESS Open Archive, 2024. 2, 10, 11, 12

    2024

  76. [76]

    WF-V AE: Enhanc- ing Video V AE by Wavelet-Driven Energy Flow for Latent Video Diffusion Model, 2025

    Zongjian Li, Bin Lin, Yang Ye, Liuhan Chen, Xinhua Cheng, Shenghai Yuan, and Li Yuan. WF-V AE: Enhanc- ing Video V AE by Wavelet-Driven Energy Flow for Latent Video Diffusion Model, 2025. 2, 3

    2025

  77. [77]

    Mahoney, Xiaoye S

    Soon Hoe Lim, Yijin Wang, Annan Yu, Emma Hart, Michael W. Mahoney, Xiaoye S. Li, and N. Benjamin Erichson. Elucidating the design choice of probability paths in flow matching for forecasting, 2024. 9

    2024

  78. [78]

    SRNDiff: Short-term Rainfall Nowcast- ing with Condition Diffusion Model, 2024

    Xudong Ling, Chaorong Li, Fengqing Qin, Peng Yang, and Yuanyuan Huang. SRNDiff: Short-term Rainfall Nowcast- ing with Condition Diffusion Model, 2024. 2, 4, 11, 12

    2024

  79. [79]

    Flow matching for generative modeling

    Yaron Lipman, Ricky TQ Chen, Heli Ben-Hamu, Maximil- ian Nickel, and Matthew Le. Flow matching for generative modeling. InICLR, 2023. 1, 3

    2023

  80. [80]

    MPL-GAN: Toward Real- istic Meteorological Predictive Learning Using Conditional GAN.IEEE Access, 8:93179–93186, 2020

    Hong-Bin Liu and Ickjai Lee. MPL-GAN: Toward Real- istic Meteorological Predictive Learning Using Conditional GAN.IEEE Access, 8:93179–93186, 2020. 1, 2, 11

    2020

Showing first 80 references.