{"paper":{"title":"PnP-Corrector: A Universal Correction Framework for Coupled Spatiotemporal Forecasting","license":"http://creativecommons.org/licenses/by/4.0/","headline":"By freezing pre-trained physics engines and training only a correction agent, the PnP-Corrector framework counters reciprocal error amplification to improve long-term accuracy in coupled spatiotemporal forecasts.","cross_cats":["cs.LG"],"primary_cat":"cs.AI","authors_text":"Fan Xu, Fan Zhang, Hao Jia, Hao Wu, Penghao Zhao, Qingsong Wen, Ruijian Gou, Xian Wu, Xiaomeng Huang, Yuan Gao, Yuxuan Liang, Yuxu Lu","submitted_at":"2026-05-09T13:12:33Z","abstract_excerpt":"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 t"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"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 29% and surpasses state-of-the-art models on several key metrics.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That a correction agent trained separately on frozen physics engines can proactively counteract systematic biases arising from reciprocal error amplification without retraining or modifying the underlying simulators.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"PnP-Corrector decouples physics simulation from error correction via a plug-and-play agent, cutting error by 29% in 300-day global ocean-atmosphere forecasts.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"By freezing pre-trained physics engines and training only a correction agent, the PnP-Corrector framework counters reciprocal error amplification to improve long-term accuracy in coupled spatiotemporal forecasts.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"d79b5eaf146473566674ada9faaa2b8b32031919409ea984a29308c0ac5f2d91"},"source":{"id":"2605.08935","kind":"arxiv","version":3},"verdict":{"id":"f434022a-3fb4-48f8-9697-972e930c19b8","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-14T21:17:58.580018Z","strongest_claim":"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 29% and surpasses state-of-the-art models on several key metrics.","one_line_summary":"PnP-Corrector decouples physics simulation from error correction via a plug-and-play agent, cutting error by 29% in 300-day global ocean-atmosphere forecasts.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That a correction agent trained separately on frozen physics engines can proactively counteract systematic biases arising from reciprocal error amplification without retraining or modifying the underlying simulators.","pith_extraction_headline":"By freezing pre-trained physics engines and training only a correction agent, the PnP-Corrector framework counters reciprocal error amplification to improve long-term accuracy in coupled spatiotemporal forecasts."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2605.08935/integrity.json","findings":[],"available":true,"detectors_run":[{"name":"claim_evidence","ran_at":"2026-05-20T08:42:01.926785Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"ai_meta_artifact","ran_at":"2026-05-19T20:41:00.667718Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"doi_title_agreement","ran_at":"2026-05-19T14:01:19.773665Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"doi_compliance","ran_at":"2026-05-19T10:40:14.687725Z","status":"completed","version":"1.0.0","findings_count":0}],"snapshot_sha256":"4c57e0ea18234353d691f535af1280cda65c14f99ff5f1303cc497d81a4c5c04"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":2,"snapshot_sha256":"909f55c61fcc3d658df6eb6a097ff65b6eb0d7275ba00e839d9f2df0fcce1114"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"}