{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:PZWHAXGZVERPDPQHTI4GFSC27M","short_pith_number":"pith:PZWHAXGZ","schema_version":"1.0","canonical_sha256":"7e6c705cd9a922f1be079a3862c85afb12abce769ccbccefcaa3f2d40b97fe9d","source":{"kind":"arxiv","id":"2303.05689","version":2},"attestation_state":"computed","paper":{"title":"Inducing Neural Collapse to a Fixed Hierarchy-Aware Frame for Reducing Mistake Severity","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Jim Davis, Tong Liang","submitted_at":"2023-03-10T03:44:01Z","abstract_excerpt":"There is a recently discovered and intriguing phenomenon called Neural Collapse: at the terminal phase of training a deep neural network for classification, the within-class penultimate feature means and the associated classifier vectors of all flat classes collapse to the vertices of a simplex Equiangular Tight Frame (ETF). Recent work has tried to exploit this phenomenon by fixing the related classifier weights to a pre-computed ETF to induce neural collapse and maximize the separation of the learned features when training with imbalanced data. In this work, we propose to fix the linear clas"},"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":"2303.05689","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.CV","submitted_at":"2023-03-10T03:44:01Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"ffc084ec314d8930fae76a104135749bc903dbe057c5e8abcd591a9c03ad92ab","abstract_canon_sha256":"3a055a986a41147093562ea1cf805e68612bff4853ded0e4f17c79d9793ec1d0"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T06:39:33.301004Z","signature_b64":"rfoDEXY9d8iTmASHbn1K37jzn67HNa5dkIFTRkQVenDT4HD9KakqQhTA2MCbLv2JacbegA7Zo+hHDWLQwq3tBA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"7e6c705cd9a922f1be079a3862c85afb12abce769ccbccefcaa3f2d40b97fe9d","last_reissued_at":"2026-07-05T06:39:33.300586Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T06:39:33.300586Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Inducing Neural Collapse to a Fixed Hierarchy-Aware Frame for Reducing Mistake Severity","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Jim Davis, Tong Liang","submitted_at":"2023-03-10T03:44:01Z","abstract_excerpt":"There is a recently discovered and intriguing phenomenon called Neural Collapse: at the terminal phase of training a deep neural network for classification, the within-class penultimate feature means and the associated classifier vectors of all flat classes collapse to the vertices of a simplex Equiangular Tight Frame (ETF). Recent work has tried to exploit this phenomenon by fixing the related classifier weights to a pre-computed ETF to induce neural collapse and maximize the separation of the learned features when training with imbalanced data. In this work, we propose to fix the linear clas"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2303.05689","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/2303.05689/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":"2303.05689","created_at":"2026-07-05T06:39:33.300641+00:00"},{"alias_kind":"arxiv_version","alias_value":"2303.05689v2","created_at":"2026-07-05T06:39:33.300641+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2303.05689","created_at":"2026-07-05T06:39:33.300641+00:00"},{"alias_kind":"pith_short_12","alias_value":"PZWHAXGZVERP","created_at":"2026-07-05T06:39:33.300641+00:00"},{"alias_kind":"pith_short_16","alias_value":"PZWHAXGZVERPDPQH","created_at":"2026-07-05T06:39:33.300641+00:00"},{"alias_kind":"pith_short_8","alias_value":"PZWHAXGZ","created_at":"2026-07-05T06:39:33.300641+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2509.02737","citing_title":"Imitate Optimal Policy: Prevail and Induce Action Collapse in Policy Gradient","ref_index":11,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/PZWHAXGZVERPDPQHTI4GFSC27M","json":"https://pith.science/pith/PZWHAXGZVERPDPQHTI4GFSC27M.json","graph_json":"https://pith.science/api/pith-number/PZWHAXGZVERPDPQHTI4GFSC27M/graph.json","events_json":"https://pith.science/api/pith-number/PZWHAXGZVERPDPQHTI4GFSC27M/events.json","paper":"https://pith.science/paper/PZWHAXGZ"},"agent_actions":{"view_html":"https://pith.science/pith/PZWHAXGZVERPDPQHTI4GFSC27M","download_json":"https://pith.science/pith/PZWHAXGZVERPDPQHTI4GFSC27M.json","view_paper":"https://pith.science/paper/PZWHAXGZ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2303.05689&json=true","fetch_graph":"https://pith.science/api/pith-number/PZWHAXGZVERPDPQHTI4GFSC27M/graph.json","fetch_events":"https://pith.science/api/pith-number/PZWHAXGZVERPDPQHTI4GFSC27M/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/PZWHAXGZVERPDPQHTI4GFSC27M/action/timestamp_anchor","attest_storage":"https://pith.science/pith/PZWHAXGZVERPDPQHTI4GFSC27M/action/storage_attestation","attest_author":"https://pith.science/pith/PZWHAXGZVERPDPQHTI4GFSC27M/action/author_attestation","sign_citation":"https://pith.science/pith/PZWHAXGZVERPDPQHTI4GFSC27M/action/citation_signature","submit_replication":"https://pith.science/pith/PZWHAXGZVERPDPQHTI4GFSC27M/action/replication_record"}},"created_at":"2026-07-05T06:39:33.300641+00:00","updated_at":"2026-07-05T06:39:33.300641+00:00"}