{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:5ZCX3LS5JVTHHC76OQ6VH3WB2Y","short_pith_number":"pith:5ZCX3LS5","schema_version":"1.0","canonical_sha256":"ee457dae5d4d66738bfe743d53eec1d63b5301258081ea33e4c2803c246db11c","source":{"kind":"arxiv","id":"2508.19830","version":2},"attestation_state":"computed","paper":{"title":"Target-Agnostic Calibration under Distribution Shift with Frequency-Aware Gradient Rectification","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Cai Xu, Wei Zhao, Yilin Zhang, You Wu, Ziyu Guan","submitted_at":"2025-08-27T12:28:26Z","abstract_excerpt":"Real-world model deployments inevitably encounter distribution shifts, rendering the confidence estimates of deep neural networks highly unreliable, posing severe risks in safety-critical applications. Existing methods improve calibration via training-time regularization or post-hoc adjustment, but often rely on access to (or simulation of) target domains, limiting practicality. We propose Frequency-aware Gradient Rectification (FGR), a target-agnostic training framework for robust calibration. From a frequency perspective, FGR applies low-pass filtering to a subset of training images to dimin"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2508.19830","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2025-08-27T12:28:26Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"e963af6e9b09793988c7f6d1b97b39cd6523644ffc6fdd2a94144a575be3c2d7","abstract_canon_sha256":"bb9fae4ef70f1338e7028d98c200d57fe2cf922a81affd40fce37f58eea8ab98"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-01T01:03:44.433285Z","signature_b64":"Memy6cb3gY1hCaNEj7XZdwEGDD/qC2Pc3FiCL+1QhbbfqnO/gdPJOGRXPGI7qkaFG7vs6BBC7Fun20saM3w4Ag==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"ee457dae5d4d66738bfe743d53eec1d63b5301258081ea33e4c2803c246db11c","last_reissued_at":"2026-06-01T01:03:44.432088Z","signature_status":"signed_v1","first_computed_at":"2026-06-01T01:03:44.432088Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Target-Agnostic Calibration under Distribution Shift with Frequency-Aware Gradient Rectification","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Cai Xu, Wei Zhao, Yilin Zhang, You Wu, Ziyu Guan","submitted_at":"2025-08-27T12:28:26Z","abstract_excerpt":"Real-world model deployments inevitably encounter distribution shifts, rendering the confidence estimates of deep neural networks highly unreliable, posing severe risks in safety-critical applications. Existing methods improve calibration via training-time regularization or post-hoc adjustment, but often rely on access to (or simulation of) target domains, limiting practicality. We propose Frequency-aware Gradient Rectification (FGR), a target-agnostic training framework for robust calibration. From a frequency perspective, FGR applies low-pass filtering to a subset of training images to dimin"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2508.19830","kind":"arxiv","version":2},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2508.19830/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"},"aliases":[{"alias_kind":"arxiv","alias_value":"2508.19830","created_at":"2026-06-01T01:03:44.432254+00:00"},{"alias_kind":"arxiv_version","alias_value":"2508.19830v2","created_at":"2026-06-01T01:03:44.432254+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2508.19830","created_at":"2026-06-01T01:03:44.432254+00:00"},{"alias_kind":"pith_short_12","alias_value":"5ZCX3LS5JVTH","created_at":"2026-06-01T01:03:44.432254+00:00"},{"alias_kind":"pith_short_16","alias_value":"5ZCX3LS5JVTHHC76","created_at":"2026-06-01T01:03:44.432254+00:00"},{"alias_kind":"pith_short_8","alias_value":"5ZCX3LS5","created_at":"2026-06-01T01:03:44.432254+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2604.12425","citing_title":"Forecasting the Past: Gradient-Based Distribution Shift Detection in Trajectory Prediction","ref_index":43,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/5ZCX3LS5JVTHHC76OQ6VH3WB2Y","json":"https://pith.science/pith/5ZCX3LS5JVTHHC76OQ6VH3WB2Y.json","graph_json":"https://pith.science/api/pith-number/5ZCX3LS5JVTHHC76OQ6VH3WB2Y/graph.json","events_json":"https://pith.science/api/pith-number/5ZCX3LS5JVTHHC76OQ6VH3WB2Y/events.json","paper":"https://pith.science/paper/5ZCX3LS5"},"agent_actions":{"view_html":"https://pith.science/pith/5ZCX3LS5JVTHHC76OQ6VH3WB2Y","download_json":"https://pith.science/pith/5ZCX3LS5JVTHHC76OQ6VH3WB2Y.json","view_paper":"https://pith.science/paper/5ZCX3LS5","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2508.19830&json=true","fetch_graph":"https://pith.science/api/pith-number/5ZCX3LS5JVTHHC76OQ6VH3WB2Y/graph.json","fetch_events":"https://pith.science/api/pith-number/5ZCX3LS5JVTHHC76OQ6VH3WB2Y/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/5ZCX3LS5JVTHHC76OQ6VH3WB2Y/action/timestamp_anchor","attest_storage":"https://pith.science/pith/5ZCX3LS5JVTHHC76OQ6VH3WB2Y/action/storage_attestation","attest_author":"https://pith.science/pith/5ZCX3LS5JVTHHC76OQ6VH3WB2Y/action/author_attestation","sign_citation":"https://pith.science/pith/5ZCX3LS5JVTHHC76OQ6VH3WB2Y/action/citation_signature","submit_replication":"https://pith.science/pith/5ZCX3LS5JVTHHC76OQ6VH3WB2Y/action/replication_record"}},"created_at":"2026-06-01T01:03:44.432254+00:00","updated_at":"2026-06-01T01:03:44.432254+00:00"}