{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:EYC4CVXMBO2HAONKXYEUFDEPRV","short_pith_number":"pith:EYC4CVXM","schema_version":"1.0","canonical_sha256":"2605c156ec0bb47039aabe09428c8f8d6a2e9f0680cf4753c28c5f979fc52d84","source":{"kind":"arxiv","id":"2402.15505","version":1},"attestation_state":"computed","paper":{"title":"Co-Supervised Learning: Improving Weak-to-Strong Generalization with Hierarchical Mixture of Experts","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.CV"],"primary_cat":"cs.LG","authors_text":"Alexandre Alahi, Yuejiang Liu","submitted_at":"2024-02-23T18:56:11Z","abstract_excerpt":"Steering the behavior of a strong model pre-trained on internet-scale data can be difficult due to the scarcity of competent supervisors. Recent studies reveal that, despite supervisory noises, a strong student model may surpass its weak teacher when fine-tuned on specific objectives. Yet, the effectiveness of such weak-to-strong generalization remains limited, especially in the presence of large capability gaps. In this paper, we propose to address this challenge by harnessing a diverse set of specialized teachers, instead of a single generalist one, that collectively supervises the strong st"},"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":"2402.15505","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-02-23T18:56:11Z","cross_cats_sorted":["cs.AI","cs.CV"],"title_canon_sha256":"2b04d712b7be3f4fffee43d37b9871e7b0f440e0e52468c8dfb1b3d773ad00b5","abstract_canon_sha256":"7f3327b85f297ba974bca898c81af4d1bb93d8fb09b485b188661e4b0235e1b5"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T07:48:41.369852Z","signature_b64":"pBZxNqGWdnReRYs+65CeEPujejyN0eDWQF2fUGknVNE0A85wQ8Pu6D8ry9pH6Tf2PPK0k9ZhgmMBK2c1FUDeCA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"2605c156ec0bb47039aabe09428c8f8d6a2e9f0680cf4753c28c5f979fc52d84","last_reissued_at":"2026-07-05T07:48:41.369427Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T07:48:41.369427Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Co-Supervised Learning: Improving Weak-to-Strong Generalization with Hierarchical Mixture of Experts","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.CV"],"primary_cat":"cs.LG","authors_text":"Alexandre Alahi, Yuejiang Liu","submitted_at":"2024-02-23T18:56:11Z","abstract_excerpt":"Steering the behavior of a strong model pre-trained on internet-scale data can be difficult due to the scarcity of competent supervisors. Recent studies reveal that, despite supervisory noises, a strong student model may surpass its weak teacher when fine-tuned on specific objectives. Yet, the effectiveness of such weak-to-strong generalization remains limited, especially in the presence of large capability gaps. In this paper, we propose to address this challenge by harnessing a diverse set of specialized teachers, instead of a single generalist one, that collectively supervises the strong st"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2402.15505","kind":"arxiv","version":1},"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/2402.15505/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":"2402.15505","created_at":"2026-07-05T07:48:41.369483+00:00"},{"alias_kind":"arxiv_version","alias_value":"2402.15505v1","created_at":"2026-07-05T07:48:41.369483+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2402.15505","created_at":"2026-07-05T07:48:41.369483+00:00"},{"alias_kind":"pith_short_12","alias_value":"EYC4CVXMBO2H","created_at":"2026-07-05T07:48:41.369483+00:00"},{"alias_kind":"pith_short_16","alias_value":"EYC4CVXMBO2HAONK","created_at":"2026-07-05T07:48:41.369483+00:00"},{"alias_kind":"pith_short_8","alias_value":"EYC4CVXM","created_at":"2026-07-05T07:48:41.369483+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":5,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.17767","citing_title":"Feature Learning in Linear-Width Two-Layer Networks: Two vs. One Step of Gradient Descent","ref_index":263,"is_internal_anchor":false},{"citing_arxiv_id":"2502.05075","citing_title":"Discrepancies are Virtue: Weak-to-Strong Generalization through Lens of Intrinsic Dimension","ref_index":15,"is_internal_anchor":false},{"citing_arxiv_id":"2505.03631","citing_title":"Generalizable Video Quality Assessment via Weak-to-Strong Learning","ref_index":9,"is_internal_anchor":false},{"citing_arxiv_id":"2605.17767","citing_title":"Feature Learning in Linear-Width Two-Layer Networks: Two vs. One Step of Gradient Descent","ref_index":263,"is_internal_anchor":false},{"citing_arxiv_id":"2605.05710","citing_title":"On the Blessing of Pre-training in Weak-to-Strong Generalization","ref_index":100,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/EYC4CVXMBO2HAONKXYEUFDEPRV","json":"https://pith.science/pith/EYC4CVXMBO2HAONKXYEUFDEPRV.json","graph_json":"https://pith.science/api/pith-number/EYC4CVXMBO2HAONKXYEUFDEPRV/graph.json","events_json":"https://pith.science/api/pith-number/EYC4CVXMBO2HAONKXYEUFDEPRV/events.json","paper":"https://pith.science/paper/EYC4CVXM"},"agent_actions":{"view_html":"https://pith.science/pith/EYC4CVXMBO2HAONKXYEUFDEPRV","download_json":"https://pith.science/pith/EYC4CVXMBO2HAONKXYEUFDEPRV.json","view_paper":"https://pith.science/paper/EYC4CVXM","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2402.15505&json=true","fetch_graph":"https://pith.science/api/pith-number/EYC4CVXMBO2HAONKXYEUFDEPRV/graph.json","fetch_events":"https://pith.science/api/pith-number/EYC4CVXMBO2HAONKXYEUFDEPRV/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/EYC4CVXMBO2HAONKXYEUFDEPRV/action/timestamp_anchor","attest_storage":"https://pith.science/pith/EYC4CVXMBO2HAONKXYEUFDEPRV/action/storage_attestation","attest_author":"https://pith.science/pith/EYC4CVXMBO2HAONKXYEUFDEPRV/action/author_attestation","sign_citation":"https://pith.science/pith/EYC4CVXMBO2HAONKXYEUFDEPRV/action/citation_signature","submit_replication":"https://pith.science/pith/EYC4CVXMBO2HAONKXYEUFDEPRV/action/replication_record"}},"created_at":"2026-07-05T07:48:41.369483+00:00","updated_at":"2026-07-05T07:48:41.369483+00:00"}