{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2022:4A4Z3SI7ADSK6CEW46ABGR4DFW","short_pith_number":"pith:4A4Z3SI7","schema_version":"1.0","canonical_sha256":"e0399dc91f00e4af0896e7801347832da01186673aa69cda7b83fab22b1384ff","source":{"kind":"arxiv","id":"2206.01278","version":1},"attestation_state":"computed","paper":{"title":"Lottery Tickets on a Data Diet: Finding Initializations with Sparse Trainable Networks","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","stat.ML"],"primary_cat":"cs.LG","authors_text":"Brett W. Larsen, Gintare Karolina Dziugaite, Jonathan Frankle, Mansheej Paul, Surya Ganguli","submitted_at":"2022-06-02T20:04:06Z","abstract_excerpt":"A striking observation about iterative magnitude pruning (IMP; Frankle et al. 2020) is that $\\unicode{x2014}$ after just a few hundred steps of dense training $\\unicode{x2014}$ the method can find a sparse sub-network that can be trained to the same accuracy as the dense network. However, the same does not hold at step 0, i.e. random initialization. In this work, we seek to understand how this early phase of pre-training leads to a good initialization for IMP both through the lens of the data distribution and the loss landscape geometry. Empirically we observe that, holding the number of pre-t"},"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":"2206.01278","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2022-06-02T20:04:06Z","cross_cats_sorted":["cs.AI","stat.ML"],"title_canon_sha256":"1aa10ff2cbf4e06b4d555d08c5f90ee19f884eef23434c4df82d9331c2660249","abstract_canon_sha256":"fba5ecd5aad05736d1286e5f73c5d0c41e93c67901d374277fc206d91788bd21"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T04:28:45.929674Z","signature_b64":"XfxVET4rYjRP6vk9cmO8BPnC+rBfCuObvMHpPnC9988VnVKy4FjfdsZjM5gbQSaz5iehoF49llmtGV6b5w2nDw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e0399dc91f00e4af0896e7801347832da01186673aa69cda7b83fab22b1384ff","last_reissued_at":"2026-07-05T04:28:45.929258Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T04:28:45.929258Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Lottery Tickets on a Data Diet: Finding Initializations with Sparse Trainable Networks","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","stat.ML"],"primary_cat":"cs.LG","authors_text":"Brett W. Larsen, Gintare Karolina Dziugaite, Jonathan Frankle, Mansheej Paul, Surya Ganguli","submitted_at":"2022-06-02T20:04:06Z","abstract_excerpt":"A striking observation about iterative magnitude pruning (IMP; Frankle et al. 2020) is that $\\unicode{x2014}$ after just a few hundred steps of dense training $\\unicode{x2014}$ the method can find a sparse sub-network that can be trained to the same accuracy as the dense network. However, the same does not hold at step 0, i.e. random initialization. In this work, we seek to understand how this early phase of pre-training leads to a good initialization for IMP both through the lens of the data distribution and the loss landscape geometry. Empirically we observe that, holding the number of pre-t"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2206.01278","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/2206.01278/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":"2206.01278","created_at":"2026-07-05T04:28:45.929320+00:00"},{"alias_kind":"arxiv_version","alias_value":"2206.01278v1","created_at":"2026-07-05T04:28:45.929320+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2206.01278","created_at":"2026-07-05T04:28:45.929320+00:00"},{"alias_kind":"pith_short_12","alias_value":"4A4Z3SI7ADSK","created_at":"2026-07-05T04:28:45.929320+00:00"},{"alias_kind":"pith_short_16","alias_value":"4A4Z3SI7ADSK6CEW","created_at":"2026-07-05T04:28:45.929320+00:00"},{"alias_kind":"pith_short_8","alias_value":"4A4Z3SI7","created_at":"2026-07-05T04:28:45.929320+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/4A4Z3SI7ADSK6CEW46ABGR4DFW","json":"https://pith.science/pith/4A4Z3SI7ADSK6CEW46ABGR4DFW.json","graph_json":"https://pith.science/api/pith-number/4A4Z3SI7ADSK6CEW46ABGR4DFW/graph.json","events_json":"https://pith.science/api/pith-number/4A4Z3SI7ADSK6CEW46ABGR4DFW/events.json","paper":"https://pith.science/paper/4A4Z3SI7"},"agent_actions":{"view_html":"https://pith.science/pith/4A4Z3SI7ADSK6CEW46ABGR4DFW","download_json":"https://pith.science/pith/4A4Z3SI7ADSK6CEW46ABGR4DFW.json","view_paper":"https://pith.science/paper/4A4Z3SI7","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2206.01278&json=true","fetch_graph":"https://pith.science/api/pith-number/4A4Z3SI7ADSK6CEW46ABGR4DFW/graph.json","fetch_events":"https://pith.science/api/pith-number/4A4Z3SI7ADSK6CEW46ABGR4DFW/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/4A4Z3SI7ADSK6CEW46ABGR4DFW/action/timestamp_anchor","attest_storage":"https://pith.science/pith/4A4Z3SI7ADSK6CEW46ABGR4DFW/action/storage_attestation","attest_author":"https://pith.science/pith/4A4Z3SI7ADSK6CEW46ABGR4DFW/action/author_attestation","sign_citation":"https://pith.science/pith/4A4Z3SI7ADSK6CEW46ABGR4DFW/action/citation_signature","submit_replication":"https://pith.science/pith/4A4Z3SI7ADSK6CEW46ABGR4DFW/action/replication_record"}},"created_at":"2026-07-05T04:28:45.929320+00:00","updated_at":"2026-07-05T04:28:45.929320+00:00"}