{"paper":{"title":"Multi-Quantile Regression for Extreme Precipitation Downscaling","license":"http://creativecommons.org/licenses/by/4.0/","headline":"A multi-quantile super-resolution network detects extreme precipitation up to 18 times better than deterministic baselines by using pinball loss on separate quantile heads.","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Gareth Lagerwall, Hamed Najafi, Jason Liu, Jayantha Obeysekera","submitted_at":"2026-05-12T21:17:26Z","abstract_excerpt":"Deep super-resolution networks for precipitation downscaling achieve strong bulk skill yet systematically under-predict the heavy-tail events that drive flood risk. We demonstrate that the primary obstacle is the loss function, not the data: under intensity-weighted MAE, real and synthetic labels at the same input are simply averaged, meaning data augmentation shifts the predicted mean rather than the conditional distribution. We resolve this with Q-SRDRN, a multi-quantile super-resolution network trained with pinball loss at tau in 0.50, 0.95, 0.99, 0.999. Two CNN-specific design choices make"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"Under this design, data augmentation via cVAE becomes complementary: the median head absorbs synthetic patterns without contaminating upper quantiles. Empirically, on Florida (convective/tropical-cyclone dominated), the un-augmented Q-SRDRN P999 head detects 1,598 of 2,111 events at 200 mm/day versus 88 for the deterministic baseline--an 18x detection-rate gain (4.2% to 75.7%)--with 63% lower KL divergence and 3.9% lower RMSE.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"The primary obstacle is the loss function, not the data: under intensity-weighted MAE, real and synthetic labels at the same input are simply averaged, meaning data augmentation shifts the predicted mean rather than the conditional distribution.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"Q-SRDRN multi-quantile network with pinball loss and per-quantile heads detects extreme precipitation events up to 18 times more effectively than deterministic baselines while preserving augmentation benefits for the median.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"A multi-quantile super-resolution network detects extreme precipitation up to 18 times better than deterministic baselines by using pinball loss on separate quantile heads.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"ac1c872a908ee5d276ed5fae9bd0ddae4cf918cfae8c8d2d85f8f88748bb7025"},"source":{"id":"2605.12762","kind":"arxiv","version":1},"verdict":{"id":"d87848b5-b7b2-443d-a758-d42f3902d6b6","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-14T21:05:01.913654Z","strongest_claim":"Under this design, data augmentation via cVAE becomes complementary: the median head absorbs synthetic patterns without contaminating upper quantiles. Empirically, on Florida (convective/tropical-cyclone dominated), the un-augmented Q-SRDRN P999 head detects 1,598 of 2,111 events at 200 mm/day versus 88 for the deterministic baseline--an 18x detection-rate gain (4.2% to 75.7%)--with 63% lower KL divergence and 3.9% lower RMSE.","one_line_summary":"Q-SRDRN multi-quantile network with pinball loss and per-quantile heads detects extreme precipitation events up to 18 times more effectively than deterministic baselines while preserving augmentation benefits for the median.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"The primary obstacle is the loss function, not the data: under intensity-weighted MAE, real and synthetic labels at the same input are simply averaged, meaning data augmentation shifts the predicted mean rather than the conditional distribution.","pith_extraction_headline":"A multi-quantile super-resolution network detects extreme precipitation up to 18 times better than deterministic baselines by using pinball loss on separate quantile heads."},"references":{"count":30,"sample":[{"doi":"10.5194/gmd-13-2109-2020","year":2020,"title":"J. Ba\\ no-Medina, R. Manzanas, and J. M. Guti \\'e rrez. Configuration and intercomparison of deep learning neural models for statistical downscaling. Geoscientific Model Development, 13(4):2109--2124,","work_id":"1e183748-4e3e-4a6d-ac2a-80e340040e53","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"10.1038/s41586-023-06185-3","year":2023,"title":"Accurate medium-range global weather forecasting with 3d neural networks","work_id":"2b852ba0-c3f8-4951-a669-5611c50c82be","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2022,"title":"A. Brando, J. Gimeno, J. A. Rodr\\'iguez-Serrano, and J. Vitri\\`a. Deep non-crossing quantiles through the partial derivative. In Proceedings of the 25th International Conference on Artificial Intellig","work_id":"6c1cf97b-cccd-4a83-84ec-cb4f9235c54d","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"10.1175/1520-0493(2004)132","year":2004,"title":"J. B. Bremnes. Probabilistic forecasts of precipitation in terms of quantiles using NWP model output. Monthly Weather Review, 132(1):338--347, 2004. DOI: 10.1175/1520-0493(2004)132<0338:PFOPIT>2.0.CO;","work_id":"9d05414e-6735-493c-afb4-86961bdd43c3","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"10.1016/j.cageo.2010.07.005","year":2011,"title":"A. J. Cannon. Quantile regression neural networks: Implementation in R and application to precipitation downscaling. Computers & Geosciences, 37(9):1277--1284, 2011. DOI: 10.1016/j.cageo.2010.07.005 h","work_id":"745f0b9d-1845-4f2f-bd84-52cf6ca2eaaf","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":30,"snapshot_sha256":"3220e8854da9df8b62e587cffd286f3b55da760bc31c9752cf74cd88412b48bb","internal_anchors":3},"formal_canon":{"evidence_count":2,"snapshot_sha256":"a661b0dfd1017d89c6d9f3f23daa9f732c8921960345504a629f505d574c3b7d"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"}