{"paper":{"title":"Reversible Residual Normalization Alleviates Spatio-Temporal Distribution Shift","license":"http://creativecommons.org/licenses/by/4.0/","headline":"Reversible Residual Normalization uses spatially-aware invertible transformations to counter distribution shifts in spatio-temporal forecasting.","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Mehdi Naima, Vincent Gauthier, Zhaobo Hu","submitted_at":"2026-04-17T08:40:28Z","abstract_excerpt":"Distribution shift severely degrades the performance of deep forecasting models. While this issue is well-studied for individual time series, it remains a significant challenge in the spatio-temporal domain. Effective solutions like instance normalization and its variants can mitigate temporal shifts by standardizing statistics. However, distribution shift on a graph is far more complex, involving not only the drift of individual node series but also heterogeneity across the spatial network where different nodes exhibit distinct statistical properties. To tackle this problem, we propose Revers"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"We propose Reversible Residual Normalization (RRN), a novel framework that performs spatially-aware invertible transformations to address distribution shift in both spatial and temporal dimensions.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That integrating graph convolutional operations inside invertible residual blocks will produce adaptive normalization that respects the underlying graph structure without introducing irreversible information loss or unstable gradients.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"Reversible Residual Normalization (RRN) introduces spatially-aware invertible residual blocks that combine center normalization with spectral-constrained graph convolutions to mitigate spatio-temporal distribution shifts in graph forecasting.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Reversible Residual Normalization uses spatially-aware invertible transformations to counter distribution shifts in spatio-temporal forecasting.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"62c75c4d05d2e2daae56b4256b9316d8148ac22cca34e5f56c74aaeba72c46d8"},"source":{"id":"2604.15838","kind":"arxiv","version":2},"verdict":{"id":"bbbda882-4c61-4a03-b823-1f09851c7d75","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-10T09:13:39.768864Z","strongest_claim":"We propose Reversible Residual Normalization (RRN), a novel framework that performs spatially-aware invertible transformations to address distribution shift in both spatial and temporal dimensions.","one_line_summary":"Reversible Residual Normalization (RRN) introduces spatially-aware invertible residual blocks that combine center normalization with spectral-constrained graph convolutions to mitigate spatio-temporal distribution shifts in graph forecasting.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That integrating graph convolutional operations inside invertible residual blocks will produce adaptive normalization that respects the underlying graph structure without introducing irreversible information loss or unstable gradients.","pith_extraction_headline":"Reversible Residual Normalization uses spatially-aware invertible transformations to counter distribution shifts in spatio-temporal forecasting."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2604.15838/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"}