PnP-Corrector: A Universal Correction Framework for Coupled Spatiotemporal Forecasting

Hao Wu, Fan Xu, Yuxu Lu, Penghao Zhao, Fan Zhang, Hao Jia, Yuxuan Liang, Ruijian Gou, Qingsong Wen, Xian Wu, Xiaomeng Huang, Yuan Gao
Proceedings of the 43rd International Conference on Machine Learning, PMLR 306:137228-137251, 2026.

Abstract

Coupled spatiotemporal forecasting is important for predicting the future evolution of multiple interacting dynamical systems, such as in climate models. However, existing methods are severely constrained by the persistent bottleneck of compounding errors. In coupled systems, errors from each subsystem simulator propagate and amplify one another, a phenomenon we term Reciprocal Error Amplification leading to a rapid collapse of long-range predictions. To address this challenge, we propose a universal framework called PnP-Corrector (Plug-and-Play Corrector). The core idea of our framework is to decouple the physical simulation from the error correction process: it freezes pre-trained physics simulation engines and exclusively trains a correction agent to proactively counteract the systematic biases emerging from the coupled system. Furthermore, we design an efficient predictive model architecture, DSLCast, to serve as the backbone of this framework. Extensive experiments demonstrate that our method significantly enhances the long-term stability and accuracy of coupled forecasting systems. For instance, in the challenging task of a 300-day global ocean-atmosphere coupled forecast, our PnP-Corrector framework reduces the prediction error of the baseline model by 28% and surpasses state-of-the-art models on several key metrics. Codes link: https://github.com/Alexander-wu/PnP-Corrector.

Cite this Paper

BibTeX
@InProceedings{pmlr-v306-wu26bo, title = {{P}n{P}-Corrector: A Universal Correction Framework for Coupled Spatiotemporal Forecasting}, author = {Wu, Hao and Xu, Fan and Lu, Yuxu and Zhao, Penghao and Zhang, Fan and Jia, Hao and Liang, Yuxuan and Gou, Ruijian and Wen, Qingsong and Wu, Xian and Huang, Xiaomeng and Gao, Yuan}, booktitle = {Proceedings of the 43rd International Conference on Machine Learning}, pages = {137228--137251}, year = {2026}, editor = {Zhang, Tong and Dudik, Miroslav and Jaggi, Martin and Agarwal, Alekh and Li, Sharon and Schuurmans, Dale and Zhu, Jerry and Berkenkamp, Felix and Dong, Hanze and Bietti, Alberto}, volume = {306}, series = {Proceedings of Machine Learning Research}, month = {06--11 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v306/main/assets/wu26bo/wu26bo.pdf}, url = {https://proceedings.mlr.press/v306/wu26bo.html}, abstract = {Coupled spatiotemporal forecasting is important for predicting the future evolution of multiple interacting dynamical systems, such as in climate models. However, existing methods are severely constrained by the persistent bottleneck of compounding errors. In coupled systems, errors from each subsystem simulator propagate and amplify one another, a phenomenon we term Reciprocal Error Amplification leading to a rapid collapse of long-range predictions. To address this challenge, we propose a universal framework called PnP-Corrector (Plug-and-Play Corrector). The core idea of our framework is to decouple the physical simulation from the error correction process: it freezes pre-trained physics simulation engines and exclusively trains a correction agent to proactively counteract the systematic biases emerging from the coupled system. Furthermore, we design an efficient predictive model architecture, DSLCast, to serve as the backbone of this framework. Extensive experiments demonstrate that our method significantly enhances the long-term stability and accuracy of coupled forecasting systems. For instance, in the challenging task of a 300-day global ocean-atmosphere coupled forecast, our PnP-Corrector framework reduces the prediction error of the baseline model by 28% and surpasses state-of-the-art models on several key metrics. Codes link: https://github.com/Alexander-wu/PnP-Corrector.} }
Endnote
%0 Conference Paper %T PnP-Corrector: A Universal Correction Framework for Coupled Spatiotemporal Forecasting %A Hao Wu %A Fan Xu %A Yuxu Lu %A Penghao Zhao %A Fan Zhang %A Hao Jia %A Yuxuan Liang %A Ruijian Gou %A Qingsong Wen %A Xian Wu %A Xiaomeng Huang %A Yuan Gao %B Proceedings of the 43rd International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2026 %E Tong Zhang %E Miroslav Dudik %E Martin Jaggi %E Alekh Agarwal %E Sharon Li %E Dale Schuurmans %E Jerry Zhu %E Felix Berkenkamp %E Hanze Dong %E Alberto Bietti %F pmlr-v306-wu26bo %I PMLR %P 137228--137251 %U https://proceedings.mlr.press/v306/wu26bo.html %V 306 %X Coupled spatiotemporal forecasting is important for predicting the future evolution of multiple interacting dynamical systems, such as in climate models. However, existing methods are severely constrained by the persistent bottleneck of compounding errors. In coupled systems, errors from each subsystem simulator propagate and amplify one another, a phenomenon we term Reciprocal Error Amplification leading to a rapid collapse of long-range predictions. To address this challenge, we propose a universal framework called PnP-Corrector (Plug-and-Play Corrector). The core idea of our framework is to decouple the physical simulation from the error correction process: it freezes pre-trained physics simulation engines and exclusively trains a correction agent to proactively counteract the systematic biases emerging from the coupled system. Furthermore, we design an efficient predictive model architecture, DSLCast, to serve as the backbone of this framework. Extensive experiments demonstrate that our method significantly enhances the long-term stability and accuracy of coupled forecasting systems. For instance, in the challenging task of a 300-day global ocean-atmosphere coupled forecast, our PnP-Corrector framework reduces the prediction error of the baseline model by 28% and surpasses state-of-the-art models on several key metrics. Codes link: https://github.com/Alexander-wu/PnP-Corrector.
APA
Wu, H., Xu, F., Lu, Y., Zhao, P., Zhang, F., Jia, H., Liang, Y., Gou, R., Wen, Q., Wu, X., Huang, X. & Gao, Y.. (2026). PnP-Corrector: A Universal Correction Framework for Coupled Spatiotemporal Forecasting. Proceedings of the 43rd International Conference on Machine Learning, in Proceedings of Machine Learning Research 306:137228-137251 Available from https://proceedings.mlr.press/v306/wu26bo.html.

Related Material