{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2020:2PCUGVD4XUSJXGKBHCGJEZZNU5","short_pith_number":"pith:2PCUGVD4","schema_version":"1.0","canonical_sha256":"d3c543547cbd249b9941388c92672da774ac336b0c7df640c1e31f3b7a3c3493","source":{"kind":"arxiv","id":"2006.12800","version":2},"attestation_state":"computed","paper":{"title":"Calibration of Neural Networks using Splines","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CV","stat.ML"],"primary_cat":"cs.LG","authors_text":"Amir Rahimi, Cristian Sminchisescu, Kartik Gupta, Richard Hartley, Thalaiyasingam Ajanthan, Thomas Mensink","submitted_at":"2020-06-23T07:18:05Z","abstract_excerpt":"Calibrating neural networks is of utmost importance when employing them in safety-critical applications where the downstream decision making depends on the predicted probabilities. Measuring calibration error amounts to comparing two empirical distributions. In this work, we introduce a binning-free calibration measure inspired by the classical Kolmogorov-Smirnov (KS) statistical test in which the main idea is to compare the respective cumulative probability distributions. From this, by approximating the empirical cumulative distribution using a differentiable function via splines, we obtain a"},"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":"2006.12800","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2020-06-23T07:18:05Z","cross_cats_sorted":["cs.CV","stat.ML"],"title_canon_sha256":"ce637cda56f68e9da602bd5cef8f04751e9d17b9d1899b3a35e07200c33d3c28","abstract_canon_sha256":"be6263b06f1b0bdd3dc9cb7bac7a5ccdc67f17dbf8e4eca088f8828998bfcc06"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T03:44:06.875823Z","signature_b64":"JYBo8Gj2ci2ngfMQGhmmbZTwQZGmFSa8uFK0D+xjkUOeer6JPd9I6o5Bv1rIVasGO7OyjPlDudG5B+dUEmDeCA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"d3c543547cbd249b9941388c92672da774ac336b0c7df640c1e31f3b7a3c3493","last_reissued_at":"2026-07-05T03:44:06.875385Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T03:44:06.875385Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Calibration of Neural Networks using Splines","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CV","stat.ML"],"primary_cat":"cs.LG","authors_text":"Amir Rahimi, Cristian Sminchisescu, Kartik Gupta, Richard Hartley, Thalaiyasingam Ajanthan, Thomas Mensink","submitted_at":"2020-06-23T07:18:05Z","abstract_excerpt":"Calibrating neural networks is of utmost importance when employing them in safety-critical applications where the downstream decision making depends on the predicted probabilities. Measuring calibration error amounts to comparing two empirical distributions. In this work, we introduce a binning-free calibration measure inspired by the classical Kolmogorov-Smirnov (KS) statistical test in which the main idea is to compare the respective cumulative probability distributions. From this, by approximating the empirical cumulative distribution using a differentiable function via splines, we obtain a"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2006.12800","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/2006.12800/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":"2006.12800","created_at":"2026-07-05T03:44:06.875441+00:00"},{"alias_kind":"arxiv_version","alias_value":"2006.12800v2","created_at":"2026-07-05T03:44:06.875441+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2006.12800","created_at":"2026-07-05T03:44:06.875441+00:00"},{"alias_kind":"pith_short_12","alias_value":"2PCUGVD4XUSJ","created_at":"2026-07-05T03:44:06.875441+00:00"},{"alias_kind":"pith_short_16","alias_value":"2PCUGVD4XUSJXGKB","created_at":"2026-07-05T03:44:06.875441+00:00"},{"alias_kind":"pith_short_8","alias_value":"2PCUGVD4","created_at":"2026-07-05T03:44:06.875441+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":4,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2607.08377","citing_title":"Eigenvalue Calibration for Semantic Embeddings of Large Language Models","ref_index":277,"is_internal_anchor":true},{"citing_arxiv_id":"2605.19369","citing_title":"When to Answer and When to Defer: A Decision Framework for Reliable Code Predictions","ref_index":5,"is_internal_anchor":false},{"citing_arxiv_id":"2605.06479","citing_title":"Risk-Controlled Post-Processing of Decision Policies","ref_index":227,"is_internal_anchor":false},{"citing_arxiv_id":"2604.08639","citing_title":"VOLTA: The Surprising Ineffectiveness of Auxiliary Losses for Calibrated Deep Learning","ref_index":5,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/2PCUGVD4XUSJXGKBHCGJEZZNU5","json":"https://pith.science/pith/2PCUGVD4XUSJXGKBHCGJEZZNU5.json","graph_json":"https://pith.science/api/pith-number/2PCUGVD4XUSJXGKBHCGJEZZNU5/graph.json","events_json":"https://pith.science/api/pith-number/2PCUGVD4XUSJXGKBHCGJEZZNU5/events.json","paper":"https://pith.science/paper/2PCUGVD4"},"agent_actions":{"view_html":"https://pith.science/pith/2PCUGVD4XUSJXGKBHCGJEZZNU5","download_json":"https://pith.science/pith/2PCUGVD4XUSJXGKBHCGJEZZNU5.json","view_paper":"https://pith.science/paper/2PCUGVD4","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2006.12800&json=true","fetch_graph":"https://pith.science/api/pith-number/2PCUGVD4XUSJXGKBHCGJEZZNU5/graph.json","fetch_events":"https://pith.science/api/pith-number/2PCUGVD4XUSJXGKBHCGJEZZNU5/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/2PCUGVD4XUSJXGKBHCGJEZZNU5/action/timestamp_anchor","attest_storage":"https://pith.science/pith/2PCUGVD4XUSJXGKBHCGJEZZNU5/action/storage_attestation","attest_author":"https://pith.science/pith/2PCUGVD4XUSJXGKBHCGJEZZNU5/action/author_attestation","sign_citation":"https://pith.science/pith/2PCUGVD4XUSJXGKBHCGJEZZNU5/action/citation_signature","submit_replication":"https://pith.science/pith/2PCUGVD4XUSJXGKBHCGJEZZNU5/action/replication_record"}},"created_at":"2026-07-05T03:44:06.875441+00:00","updated_at":"2026-07-05T03:44:06.875441+00:00"}