{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:DN24DYZUUPM3CRVQCMWRKIPJN5","short_pith_number":"pith:DN24DYZU","schema_version":"1.0","canonical_sha256":"1b75c1e334a3d9b146b0132d1521e96f7b3d504713740f64afc06038254dd4f4","source":{"kind":"arxiv","id":"2302.12091","version":2},"attestation_state":"computed","paper":{"title":"Random Teachers are Good Teachers","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Felix Sarnthein, Gregor Bachmann, Sotiris Anagnostidis, Thomas Hofmann","submitted_at":"2023-02-23T15:26:08Z","abstract_excerpt":"In this work, we investigate the implicit regularization induced by teacher-student learning dynamics in self-distillation. To isolate its effect, we describe a simple experiment where we consider teachers at random initialization instead of trained teachers. Surprisingly, when distilling a student into such a random teacher, we observe that the resulting model and its representations already possess very interesting characteristics; (1) we observe a strong improvement of the distilled student over its teacher in terms of probing accuracy. (2) The learned representations are data-dependent and"},"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":"2302.12091","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2023-02-23T15:26:08Z","cross_cats_sorted":[],"title_canon_sha256":"c1421c815d57f98603f938e5f1e3c38d7c6ed6a8ae1d15ae47eb8b76d2da151b","abstract_canon_sha256":"d6ab60e1ed9cb2fc93ab8c31d150325ebf7b5f7d9ad12ef5638e4ec503665853"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T06:27:46.322285Z","signature_b64":"Q2O2uPT5Y3XtWWdrqu651EX2Szb1NY/432dAGa9HwerY6uSRpXBpukAd6ovgARjWq6gaUMFnwSZhm5E5TWbrAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"1b75c1e334a3d9b146b0132d1521e96f7b3d504713740f64afc06038254dd4f4","last_reissued_at":"2026-07-05T06:27:46.321693Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T06:27:46.321693Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Random Teachers are Good Teachers","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Felix Sarnthein, Gregor Bachmann, Sotiris Anagnostidis, Thomas Hofmann","submitted_at":"2023-02-23T15:26:08Z","abstract_excerpt":"In this work, we investigate the implicit regularization induced by teacher-student learning dynamics in self-distillation. To isolate its effect, we describe a simple experiment where we consider teachers at random initialization instead of trained teachers. Surprisingly, when distilling a student into such a random teacher, we observe that the resulting model and its representations already possess very interesting characteristics; (1) we observe a strong improvement of the distilled student over its teacher in terms of probing accuracy. (2) The learned representations are data-dependent and"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2302.12091","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/2302.12091/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":"2302.12091","created_at":"2026-07-05T06:27:46.321774+00:00"},{"alias_kind":"arxiv_version","alias_value":"2302.12091v2","created_at":"2026-07-05T06:27:46.321774+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2302.12091","created_at":"2026-07-05T06:27:46.321774+00:00"},{"alias_kind":"pith_short_12","alias_value":"DN24DYZUUPM3","created_at":"2026-07-05T06:27:46.321774+00:00"},{"alias_kind":"pith_short_16","alias_value":"DN24DYZUUPM3CRVQ","created_at":"2026-07-05T06:27:46.321774+00:00"},{"alias_kind":"pith_short_8","alias_value":"DN24DYZU","created_at":"2026-07-05T06:27:46.321774+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2604.18390","citing_title":"Randomly Initialized Networks Can Learn from Peer-to-Peer Consensus","ref_index":11,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/DN24DYZUUPM3CRVQCMWRKIPJN5","json":"https://pith.science/pith/DN24DYZUUPM3CRVQCMWRKIPJN5.json","graph_json":"https://pith.science/api/pith-number/DN24DYZUUPM3CRVQCMWRKIPJN5/graph.json","events_json":"https://pith.science/api/pith-number/DN24DYZUUPM3CRVQCMWRKIPJN5/events.json","paper":"https://pith.science/paper/DN24DYZU"},"agent_actions":{"view_html":"https://pith.science/pith/DN24DYZUUPM3CRVQCMWRKIPJN5","download_json":"https://pith.science/pith/DN24DYZUUPM3CRVQCMWRKIPJN5.json","view_paper":"https://pith.science/paper/DN24DYZU","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2302.12091&json=true","fetch_graph":"https://pith.science/api/pith-number/DN24DYZUUPM3CRVQCMWRKIPJN5/graph.json","fetch_events":"https://pith.science/api/pith-number/DN24DYZUUPM3CRVQCMWRKIPJN5/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/DN24DYZUUPM3CRVQCMWRKIPJN5/action/timestamp_anchor","attest_storage":"https://pith.science/pith/DN24DYZUUPM3CRVQCMWRKIPJN5/action/storage_attestation","attest_author":"https://pith.science/pith/DN24DYZUUPM3CRVQCMWRKIPJN5/action/author_attestation","sign_citation":"https://pith.science/pith/DN24DYZUUPM3CRVQCMWRKIPJN5/action/citation_signature","submit_replication":"https://pith.science/pith/DN24DYZUUPM3CRVQCMWRKIPJN5/action/replication_record"}},"created_at":"2026-07-05T06:27:46.321774+00:00","updated_at":"2026-07-05T06:27:46.321774+00:00"}