{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:LOQ6H2ANCTGOPDMVXLC4452RXO","short_pith_number":"pith:LOQ6H2AN","schema_version":"1.0","canonical_sha256":"5ba1e3e80d14cce78d95bac5ce7751bbb0be7d3213bd157fa74b0d98affbb996","source":{"kind":"arxiv","id":"2106.11642","version":3},"attestation_state":"computed","paper":{"title":"Repulsive Deep Ensembles are Bayesian","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["stat.ML"],"primary_cat":"cs.LG","authors_text":"Francesco D'Angelo, Vincent Fortuin","submitted_at":"2021-06-22T09:50:28Z","abstract_excerpt":"Deep ensembles have recently gained popularity in the deep learning community for their conceptual simplicity and efficiency. However, maintaining functional diversity between ensemble members that are independently trained with gradient descent is challenging. This can lead to pathologies when adding more ensemble members, such as a saturation of the ensemble performance, which converges to the performance of a single model. Moreover, this does not only affect the quality of its predictions, but even more so the uncertainty estimates of the ensemble, and thus its performance on out-of-distrib"},"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":"2106.11642","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2021-06-22T09:50:28Z","cross_cats_sorted":["stat.ML"],"title_canon_sha256":"d9ec9597928b24ad8cf3809bb01b7928bcd4d3250c0aaa107e75fde0242b1776","abstract_canon_sha256":"c434eed72e330767341f292872ec7eabe17f400c510b7476bb7e4a5f0c1f47b0"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T05:55:06.491673Z","signature_b64":"OY4U/PkJaPXNc8qByxKlDFxEFKoNx7Y/OyNeYiaHR4KV1LJwUAVTYe67XmGakRYW5ZrVjcobfS2+mfR/CY1NAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"5ba1e3e80d14cce78d95bac5ce7751bbb0be7d3213bd157fa74b0d98affbb996","last_reissued_at":"2026-07-05T05:55:06.491188Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T05:55:06.491188Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Repulsive Deep Ensembles are Bayesian","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["stat.ML"],"primary_cat":"cs.LG","authors_text":"Francesco D'Angelo, Vincent Fortuin","submitted_at":"2021-06-22T09:50:28Z","abstract_excerpt":"Deep ensembles have recently gained popularity in the deep learning community for their conceptual simplicity and efficiency. However, maintaining functional diversity between ensemble members that are independently trained with gradient descent is challenging. This can lead to pathologies when adding more ensemble members, such as a saturation of the ensemble performance, which converges to the performance of a single model. Moreover, this does not only affect the quality of its predictions, but even more so the uncertainty estimates of the ensemble, and thus its performance on out-of-distrib"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2106.11642","kind":"arxiv","version":3},"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/2106.11642/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":"2106.11642","created_at":"2026-07-05T05:55:06.491244+00:00"},{"alias_kind":"arxiv_version","alias_value":"2106.11642v3","created_at":"2026-07-05T05:55:06.491244+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2106.11642","created_at":"2026-07-05T05:55:06.491244+00:00"},{"alias_kind":"pith_short_12","alias_value":"LOQ6H2ANCTGO","created_at":"2026-07-05T05:55:06.491244+00:00"},{"alias_kind":"pith_short_16","alias_value":"LOQ6H2ANCTGOPDMV","created_at":"2026-07-05T05:55:06.491244+00:00"},{"alias_kind":"pith_short_8","alias_value":"LOQ6H2AN","created_at":"2026-07-05T05:55:06.491244+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2509.00155","citing_title":"Amplitude Uncertainties Everywhere All at Once","ref_index":62,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/LOQ6H2ANCTGOPDMVXLC4452RXO","json":"https://pith.science/pith/LOQ6H2ANCTGOPDMVXLC4452RXO.json","graph_json":"https://pith.science/api/pith-number/LOQ6H2ANCTGOPDMVXLC4452RXO/graph.json","events_json":"https://pith.science/api/pith-number/LOQ6H2ANCTGOPDMVXLC4452RXO/events.json","paper":"https://pith.science/paper/LOQ6H2AN"},"agent_actions":{"view_html":"https://pith.science/pith/LOQ6H2ANCTGOPDMVXLC4452RXO","download_json":"https://pith.science/pith/LOQ6H2ANCTGOPDMVXLC4452RXO.json","view_paper":"https://pith.science/paper/LOQ6H2AN","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2106.11642&json=true","fetch_graph":"https://pith.science/api/pith-number/LOQ6H2ANCTGOPDMVXLC4452RXO/graph.json","fetch_events":"https://pith.science/api/pith-number/LOQ6H2ANCTGOPDMVXLC4452RXO/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/LOQ6H2ANCTGOPDMVXLC4452RXO/action/timestamp_anchor","attest_storage":"https://pith.science/pith/LOQ6H2ANCTGOPDMVXLC4452RXO/action/storage_attestation","attest_author":"https://pith.science/pith/LOQ6H2ANCTGOPDMVXLC4452RXO/action/author_attestation","sign_citation":"https://pith.science/pith/LOQ6H2ANCTGOPDMVXLC4452RXO/action/citation_signature","submit_replication":"https://pith.science/pith/LOQ6H2ANCTGOPDMVXLC4452RXO/action/replication_record"}},"created_at":"2026-07-05T05:55:06.491244+00:00","updated_at":"2026-07-05T05:55:06.491244+00:00"}