{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:VJBOKZZ4CI4ZOCBAH36EB7NAU7","short_pith_number":"pith:VJBOKZZ4","schema_version":"1.0","canonical_sha256":"aa42e5673c12399708203efc40fda0a7c6e4a96731441bd9ff43ba0d09887f03","source":{"kind":"arxiv","id":"2102.01670","version":2},"attestation_state":"computed","paper":{"title":"Keep the Gradients Flowing: Using Gradient Flow to Study Sparse Network Optimization","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CV"],"primary_cat":"cs.LG","authors_text":"Benjamin Rosman, Kale-ab Tessera, Sara Hooker","submitted_at":"2021-02-02T18:40:26Z","abstract_excerpt":"Training sparse networks to converge to the same performance as dense neural architectures has proven to be elusive. Recent work suggests that initialization is the key. However, while this direction of research has had some success, focusing on initialization alone appears to be inadequate. In this paper, we take a broader view of training sparse networks and consider the role of regularization, optimization, and architecture choices on sparse models. We propose a simple experimental framework, Same Capacity Sparse vs Dense Comparison (SC-SDC), that allows for a fair comparison of sparse 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":"2102.01670","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2021-02-02T18:40:26Z","cross_cats_sorted":["cs.CV"],"title_canon_sha256":"979391ea6827db827c092f487458bda7de9929a9c117097db0613aafd23986f1","abstract_canon_sha256":"98972d5ddcab84c02aff9964d17a67aac8b1e43ffd89d71b4f26d48e592a143a"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T02:49:53.022038Z","signature_b64":"6QCY2xTkaVoJZSUcBfqk9Ypmi/Drf5i8JlmBq6ZWc7F/eS05k/p2HEX3vJHFJhLMKm/uyCmjA0vzySLTyX/aBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"aa42e5673c12399708203efc40fda0a7c6e4a96731441bd9ff43ba0d09887f03","last_reissued_at":"2026-07-05T02:49:53.021619Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T02:49:53.021619Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Keep the Gradients Flowing: Using Gradient Flow to Study Sparse Network Optimization","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CV"],"primary_cat":"cs.LG","authors_text":"Benjamin Rosman, Kale-ab Tessera, Sara Hooker","submitted_at":"2021-02-02T18:40:26Z","abstract_excerpt":"Training sparse networks to converge to the same performance as dense neural architectures has proven to be elusive. Recent work suggests that initialization is the key. However, while this direction of research has had some success, focusing on initialization alone appears to be inadequate. In this paper, we take a broader view of training sparse networks and consider the role of regularization, optimization, and architecture choices on sparse models. We propose a simple experimental framework, Same Capacity Sparse vs Dense Comparison (SC-SDC), that allows for a fair comparison of sparse and "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2102.01670","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/2102.01670/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":"2102.01670","created_at":"2026-07-05T02:49:53.021674+00:00"},{"alias_kind":"arxiv_version","alias_value":"2102.01670v2","created_at":"2026-07-05T02:49:53.021674+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2102.01670","created_at":"2026-07-05T02:49:53.021674+00:00"},{"alias_kind":"pith_short_12","alias_value":"VJBOKZZ4CI4Z","created_at":"2026-07-05T02:49:53.021674+00:00"},{"alias_kind":"pith_short_16","alias_value":"VJBOKZZ4CI4ZOCBA","created_at":"2026-07-05T02:49:53.021674+00:00"},{"alias_kind":"pith_short_8","alias_value":"VJBOKZZ4","created_at":"2026-07-05T02:49:53.021674+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/VJBOKZZ4CI4ZOCBAH36EB7NAU7","json":"https://pith.science/pith/VJBOKZZ4CI4ZOCBAH36EB7NAU7.json","graph_json":"https://pith.science/api/pith-number/VJBOKZZ4CI4ZOCBAH36EB7NAU7/graph.json","events_json":"https://pith.science/api/pith-number/VJBOKZZ4CI4ZOCBAH36EB7NAU7/events.json","paper":"https://pith.science/paper/VJBOKZZ4"},"agent_actions":{"view_html":"https://pith.science/pith/VJBOKZZ4CI4ZOCBAH36EB7NAU7","download_json":"https://pith.science/pith/VJBOKZZ4CI4ZOCBAH36EB7NAU7.json","view_paper":"https://pith.science/paper/VJBOKZZ4","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2102.01670&json=true","fetch_graph":"https://pith.science/api/pith-number/VJBOKZZ4CI4ZOCBAH36EB7NAU7/graph.json","fetch_events":"https://pith.science/api/pith-number/VJBOKZZ4CI4ZOCBAH36EB7NAU7/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/VJBOKZZ4CI4ZOCBAH36EB7NAU7/action/timestamp_anchor","attest_storage":"https://pith.science/pith/VJBOKZZ4CI4ZOCBAH36EB7NAU7/action/storage_attestation","attest_author":"https://pith.science/pith/VJBOKZZ4CI4ZOCBAH36EB7NAU7/action/author_attestation","sign_citation":"https://pith.science/pith/VJBOKZZ4CI4ZOCBAH36EB7NAU7/action/citation_signature","submit_replication":"https://pith.science/pith/VJBOKZZ4CI4ZOCBAH36EB7NAU7/action/replication_record"}},"created_at":"2026-07-05T02:49:53.021674+00:00","updated_at":"2026-07-05T02:49:53.021674+00:00"}