{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:CCCTYZXVBI7P3VIUVWKRNZSMDI","short_pith_number":"pith:CCCTYZXV","schema_version":"1.0","canonical_sha256":"10853c66f50a3efdd514ad9516e64c1a17ebdd9b2d55ef0614cc663c331a113f","source":{"kind":"arxiv","id":"2102.02887","version":3},"attestation_state":"computed","paper":{"title":"Do We Actually Need Dense Over-Parameterization? In-Time Over-Parameterization in Sparse Training","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.CV"],"primary_cat":"cs.LG","authors_text":"Decebal Constantin Mocanu, Lu Yin, Mykola Pechenizkiy, Shiwei Liu","submitted_at":"2021-02-04T20:59:31Z","abstract_excerpt":"In this paper, we introduce a new perspective on training deep neural networks capable of state-of-the-art performance without the need for the expensive over-parameterization by proposing the concept of In-Time Over-Parameterization (ITOP) in sparse training. By starting from a random sparse network and continuously exploring sparse connectivities during training, we can perform an Over-Parameterization in the space-time manifold, closing the gap in the expressibility between sparse training and dense training. We further use ITOP to understand the underlying mechanism of Dynamic Sparse Train"},"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":"2102.02887","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2021-02-04T20:59:31Z","cross_cats_sorted":["cs.AI","cs.CV"],"title_canon_sha256":"3b1ba7063e72328bd4dcf5eb7ee563f0390c3cb308f6288abc7a74a0d919b7dc","abstract_canon_sha256":"c2f0f1c8e7e6b20dd26430756921fac760d31394f6c51ef798471803ff567515"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T02:49:24.609398Z","signature_b64":"WlHI49BZIpIBETOFDF+4pAD6nLOEs4y30/FJ6c2h4Utw288bcoTzuNRudZUK6h9EIdmZlscvi9t84Jua+WeoDg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"10853c66f50a3efdd514ad9516e64c1a17ebdd9b2d55ef0614cc663c331a113f","last_reissued_at":"2026-07-05T02:49:24.608975Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T02:49:24.608975Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Do We Actually Need Dense Over-Parameterization? In-Time Over-Parameterization in Sparse Training","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.CV"],"primary_cat":"cs.LG","authors_text":"Decebal Constantin Mocanu, Lu Yin, Mykola Pechenizkiy, Shiwei Liu","submitted_at":"2021-02-04T20:59:31Z","abstract_excerpt":"In this paper, we introduce a new perspective on training deep neural networks capable of state-of-the-art performance without the need for the expensive over-parameterization by proposing the concept of In-Time Over-Parameterization (ITOP) in sparse training. By starting from a random sparse network and continuously exploring sparse connectivities during training, we can perform an Over-Parameterization in the space-time manifold, closing the gap in the expressibility between sparse training and dense training. We further use ITOP to understand the underlying mechanism of Dynamic Sparse Train"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2102.02887","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/2102.02887/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":"2102.02887","created_at":"2026-07-05T02:49:24.609029+00:00"},{"alias_kind":"arxiv_version","alias_value":"2102.02887v3","created_at":"2026-07-05T02:49:24.609029+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2102.02887","created_at":"2026-07-05T02:49:24.609029+00:00"},{"alias_kind":"pith_short_12","alias_value":"CCCTYZXVBI7P","created_at":"2026-07-05T02:49:24.609029+00:00"},{"alias_kind":"pith_short_16","alias_value":"CCCTYZXVBI7P3VIU","created_at":"2026-07-05T02:49:24.609029+00:00"},{"alias_kind":"pith_short_8","alias_value":"CCCTYZXV","created_at":"2026-07-05T02:49:24.609029+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2310.02277","citing_title":"Junk DNA Hypothesis: Pruning Small Pre-Trained Weights Irreversibly and Monotonically Impairs \"Difficult\" Downstream Tasks in LLMs","ref_index":31,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/CCCTYZXVBI7P3VIUVWKRNZSMDI","json":"https://pith.science/pith/CCCTYZXVBI7P3VIUVWKRNZSMDI.json","graph_json":"https://pith.science/api/pith-number/CCCTYZXVBI7P3VIUVWKRNZSMDI/graph.json","events_json":"https://pith.science/api/pith-number/CCCTYZXVBI7P3VIUVWKRNZSMDI/events.json","paper":"https://pith.science/paper/CCCTYZXV"},"agent_actions":{"view_html":"https://pith.science/pith/CCCTYZXVBI7P3VIUVWKRNZSMDI","download_json":"https://pith.science/pith/CCCTYZXVBI7P3VIUVWKRNZSMDI.json","view_paper":"https://pith.science/paper/CCCTYZXV","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2102.02887&json=true","fetch_graph":"https://pith.science/api/pith-number/CCCTYZXVBI7P3VIUVWKRNZSMDI/graph.json","fetch_events":"https://pith.science/api/pith-number/CCCTYZXVBI7P3VIUVWKRNZSMDI/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/CCCTYZXVBI7P3VIUVWKRNZSMDI/action/timestamp_anchor","attest_storage":"https://pith.science/pith/CCCTYZXVBI7P3VIUVWKRNZSMDI/action/storage_attestation","attest_author":"https://pith.science/pith/CCCTYZXVBI7P3VIUVWKRNZSMDI/action/author_attestation","sign_citation":"https://pith.science/pith/CCCTYZXVBI7P3VIUVWKRNZSMDI/action/citation_signature","submit_replication":"https://pith.science/pith/CCCTYZXVBI7P3VIUVWKRNZSMDI/action/replication_record"}},"created_at":"2026-07-05T02:49:24.609029+00:00","updated_at":"2026-07-05T02:49:24.609029+00:00"}