{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2020:T62PC5EHK554V7WYPGZKDXLFOP","short_pith_number":"pith:T62PC5EH","schema_version":"1.0","canonical_sha256":"9fb4f17487577bcafed879b2a1dd6573e4447c2a069dc726298ebae3cef9c896","source":{"kind":"arxiv","id":"2012.03107","version":3},"attestation_state":"computed","paper":{"title":"When Do Curricula Work?","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CV","eess.IV","stat.ML"],"primary_cat":"cs.LG","authors_text":"Behnam Neyshabur, Ethan Dyer, Xiaoxia Wu","submitted_at":"2020-12-05T19:41:30Z","abstract_excerpt":"Inspired by human learning, researchers have proposed ordering examples during training based on their difficulty. Both curriculum learning, exposing a network to easier examples early in training, and anti-curriculum learning, showing the most difficult examples first, have been suggested as improvements to the standard i.i.d. training. In this work, we set out to investigate the relative benefits of ordered learning. We first investigate the \\emph{implicit curricula} resulting from architectural and optimization bias and find that samples are learned in a highly consistent order. Next, to qu"},"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":"2012.03107","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2020-12-05T19:41:30Z","cross_cats_sorted":["cs.CV","eess.IV","stat.ML"],"title_canon_sha256":"f501645c89f723beb19cc7bbe0edecb8b75a5227226da411120b16d9ef11633d","abstract_canon_sha256":"b3fbdff618b0d5a21a69a7b704c49e29421f4dc0c5cd0f0ca1b2763071f22897"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T02:13:56.259482Z","signature_b64":"EJ+UBSS+Q0FQ3za9se44q2dQ1vajdckvfKpiqXUAlj8oFHzLXbOsBtXjFzDx9AVnvH9ja3eslKxeWKK3JCmFAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"9fb4f17487577bcafed879b2a1dd6573e4447c2a069dc726298ebae3cef9c896","last_reissued_at":"2026-07-05T02:13:56.258815Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T02:13:56.258815Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"When Do Curricula Work?","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CV","eess.IV","stat.ML"],"primary_cat":"cs.LG","authors_text":"Behnam Neyshabur, Ethan Dyer, Xiaoxia Wu","submitted_at":"2020-12-05T19:41:30Z","abstract_excerpt":"Inspired by human learning, researchers have proposed ordering examples during training based on their difficulty. Both curriculum learning, exposing a network to easier examples early in training, and anti-curriculum learning, showing the most difficult examples first, have been suggested as improvements to the standard i.i.d. training. In this work, we set out to investigate the relative benefits of ordered learning. We first investigate the \\emph{implicit curricula} resulting from architectural and optimization bias and find that samples are learned in a highly consistent order. Next, to qu"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2012.03107","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/2012.03107/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":"2012.03107","created_at":"2026-07-05T02:13:56.258876+00:00"},{"alias_kind":"arxiv_version","alias_value":"2012.03107v3","created_at":"2026-07-05T02:13:56.258876+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2012.03107","created_at":"2026-07-05T02:13:56.258876+00:00"},{"alias_kind":"pith_short_12","alias_value":"T62PC5EHK554","created_at":"2026-07-05T02:13:56.258876+00:00"},{"alias_kind":"pith_short_16","alias_value":"T62PC5EHK554V7WY","created_at":"2026-07-05T02:13:56.258876+00:00"},{"alias_kind":"pith_short_8","alias_value":"T62PC5EH","created_at":"2026-07-05T02:13:56.258876+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":5,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.17706","citing_title":"Confusion-Aware Transfer Teacher Curriculum Learning Framework: Disentangling Scoring and Pacing Effects","ref_index":13,"is_internal_anchor":false},{"citing_arxiv_id":"2603.22586","citing_title":"A Foundation Model for Instruction-Conditioned In-Context Time Series Tasks","ref_index":20,"is_internal_anchor":false},{"citing_arxiv_id":"2603.22586","citing_title":"A Foundation Model for Instruction-Conditioned In-Context Time Series Tasks","ref_index":20,"is_internal_anchor":false},{"citing_arxiv_id":"2605.11260","citing_title":"Curriculum Learning-Guided Progressive Distillation in Large Language Models","ref_index":18,"is_internal_anchor":false},{"citing_arxiv_id":"2604.24171","citing_title":"POCA: Pareto-Optimal Curriculum Alignment for Visual Text Generation","ref_index":41,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/T62PC5EHK554V7WYPGZKDXLFOP","json":"https://pith.science/pith/T62PC5EHK554V7WYPGZKDXLFOP.json","graph_json":"https://pith.science/api/pith-number/T62PC5EHK554V7WYPGZKDXLFOP/graph.json","events_json":"https://pith.science/api/pith-number/T62PC5EHK554V7WYPGZKDXLFOP/events.json","paper":"https://pith.science/paper/T62PC5EH"},"agent_actions":{"view_html":"https://pith.science/pith/T62PC5EHK554V7WYPGZKDXLFOP","download_json":"https://pith.science/pith/T62PC5EHK554V7WYPGZKDXLFOP.json","view_paper":"https://pith.science/paper/T62PC5EH","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2012.03107&json=true","fetch_graph":"https://pith.science/api/pith-number/T62PC5EHK554V7WYPGZKDXLFOP/graph.json","fetch_events":"https://pith.science/api/pith-number/T62PC5EHK554V7WYPGZKDXLFOP/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/T62PC5EHK554V7WYPGZKDXLFOP/action/timestamp_anchor","attest_storage":"https://pith.science/pith/T62PC5EHK554V7WYPGZKDXLFOP/action/storage_attestation","attest_author":"https://pith.science/pith/T62PC5EHK554V7WYPGZKDXLFOP/action/author_attestation","sign_citation":"https://pith.science/pith/T62PC5EHK554V7WYPGZKDXLFOP/action/citation_signature","submit_replication":"https://pith.science/pith/T62PC5EHK554V7WYPGZKDXLFOP/action/replication_record"}},"created_at":"2026-07-05T02:13:56.258876+00:00","updated_at":"2026-07-05T02:13:56.258876+00:00"}