{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:C7T7DD4MLCVMRCWCIQX75N5HGE","short_pith_number":"pith:C7T7DD4M","schema_version":"1.0","canonical_sha256":"17e7f18f8c58aac88ac2442ffeb7a7312c30fc277dadb52a7b488f7024b1f883","source":{"kind":"arxiv","id":"2409.18265","version":2},"attestation_state":"computed","paper":{"title":"Task-recency bias strikes back: Adapting covariances in Exemplar-Free Class Incremental Learning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CV"],"primary_cat":"cs.LG","authors_text":"Bart{\\l}omiej Twardowski, Grzegorz Rype\\'s\\'c, Sebastian Cygert, Tomasz Trzci\\'nski","submitted_at":"2024-09-26T20:18:14Z","abstract_excerpt":"Exemplar-Free Class Incremental Learning (EFCIL) tackles the problem of training a model on a sequence of tasks without access to past data. Existing state-of-the-art methods represent classes as Gaussian distributions in the feature extractor's latent space, enabling Bayes classification or training the classifier by replaying pseudo features. However, we identify two critical issues that compromise their efficacy when the feature extractor is updated on incremental tasks. First, they do not consider that classes' covariance matrices change and must be adapted after each task. Second, they ar"},"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":"2409.18265","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-09-26T20:18:14Z","cross_cats_sorted":["cs.CV"],"title_canon_sha256":"da54c255a167aee6f62e6a44ad2d7bce8d33ad128cc7d01a390656fb2098a8d3","abstract_canon_sha256":"82540dfea652b9cdaeba85cce25d72c7380b9ed55938d1fa84bee58e74608f06"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:26:30.766922Z","signature_b64":"WoJWOXHn/lqTYC538NafHKXLnqDvI1520crLurcvq33b4vBy4Grxx4i3hetyS9nSoYDL2A52HxxSXdvryoRLBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"17e7f18f8c58aac88ac2442ffeb7a7312c30fc277dadb52a7b488f7024b1f883","last_reissued_at":"2026-07-05T09:26:30.766464Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:26:30.766464Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Task-recency bias strikes back: Adapting covariances in Exemplar-Free Class Incremental Learning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CV"],"primary_cat":"cs.LG","authors_text":"Bart{\\l}omiej Twardowski, Grzegorz Rype\\'s\\'c, Sebastian Cygert, Tomasz Trzci\\'nski","submitted_at":"2024-09-26T20:18:14Z","abstract_excerpt":"Exemplar-Free Class Incremental Learning (EFCIL) tackles the problem of training a model on a sequence of tasks without access to past data. Existing state-of-the-art methods represent classes as Gaussian distributions in the feature extractor's latent space, enabling Bayes classification or training the classifier by replaying pseudo features. However, we identify two critical issues that compromise their efficacy when the feature extractor is updated on incremental tasks. First, they do not consider that classes' covariance matrices change and must be adapted after each task. Second, they ar"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2409.18265","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/2409.18265/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":"2409.18265","created_at":"2026-07-05T09:26:30.766522+00:00"},{"alias_kind":"arxiv_version","alias_value":"2409.18265v2","created_at":"2026-07-05T09:26:30.766522+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2409.18265","created_at":"2026-07-05T09:26:30.766522+00:00"},{"alias_kind":"pith_short_12","alias_value":"C7T7DD4MLCVM","created_at":"2026-07-05T09:26:30.766522+00:00"},{"alias_kind":"pith_short_16","alias_value":"C7T7DD4MLCVMRCWC","created_at":"2026-07-05T09:26:30.766522+00:00"},{"alias_kind":"pith_short_8","alias_value":"C7T7DD4M","created_at":"2026-07-05T09:26:30.766522+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/C7T7DD4MLCVMRCWCIQX75N5HGE","json":"https://pith.science/pith/C7T7DD4MLCVMRCWCIQX75N5HGE.json","graph_json":"https://pith.science/api/pith-number/C7T7DD4MLCVMRCWCIQX75N5HGE/graph.json","events_json":"https://pith.science/api/pith-number/C7T7DD4MLCVMRCWCIQX75N5HGE/events.json","paper":"https://pith.science/paper/C7T7DD4M"},"agent_actions":{"view_html":"https://pith.science/pith/C7T7DD4MLCVMRCWCIQX75N5HGE","download_json":"https://pith.science/pith/C7T7DD4MLCVMRCWCIQX75N5HGE.json","view_paper":"https://pith.science/paper/C7T7DD4M","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2409.18265&json=true","fetch_graph":"https://pith.science/api/pith-number/C7T7DD4MLCVMRCWCIQX75N5HGE/graph.json","fetch_events":"https://pith.science/api/pith-number/C7T7DD4MLCVMRCWCIQX75N5HGE/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/C7T7DD4MLCVMRCWCIQX75N5HGE/action/timestamp_anchor","attest_storage":"https://pith.science/pith/C7T7DD4MLCVMRCWCIQX75N5HGE/action/storage_attestation","attest_author":"https://pith.science/pith/C7T7DD4MLCVMRCWCIQX75N5HGE/action/author_attestation","sign_citation":"https://pith.science/pith/C7T7DD4MLCVMRCWCIQX75N5HGE/action/citation_signature","submit_replication":"https://pith.science/pith/C7T7DD4MLCVMRCWCIQX75N5HGE/action/replication_record"}},"created_at":"2026-07-05T09:26:30.766522+00:00","updated_at":"2026-07-05T09:26:30.766522+00:00"}