{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:J3Y3OBQIFQY2EMGJEAZGXGSJQD","short_pith_number":"pith:J3Y3OBQI","schema_version":"1.0","canonical_sha256":"4ef1b706082c31a230c920326b9a4980ebfd6a9f0a0f7602cd70c6a904a1c7b7","source":{"kind":"arxiv","id":"2406.08658","version":1},"attestation_state":"computed","paper":{"title":"Pruning is Optimal for Learning Sparse Features in High-Dimensions","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"stat.ML","authors_text":"Murat A. Erdogdu, Nuri Mert Vural","submitted_at":"2024-06-12T21:43:12Z","abstract_excerpt":"While it is commonly observed in practice that pruning networks to a certain level of sparsity can improve the quality of the features, a theoretical explanation of this phenomenon remains elusive. In this work, we investigate this by demonstrating that a broad class of statistical models can be optimally learned using pruned neural networks trained with gradient descent, in high-dimensions.\n  We consider learning both single-index and multi-index models of the form $y = \\sigma^*(\\boldsymbol{V}^{\\top} \\boldsymbol{x}) + \\epsilon$, where $\\sigma^*$ is a degree-$p$ polynomial, and $\\boldsymbol{V}"},"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":"2406.08658","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"stat.ML","submitted_at":"2024-06-12T21:43:12Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"38426dd385f767a4a79fad1ce43217873dadb68fa3469a456b88b526abcfc165","abstract_canon_sha256":"1defba35ad07627112001bcdcb2aab154ba30341913169405f5eb62a555cb596"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:31:20.886512Z","signature_b64":"FZnYQt5ZGXcEWjD2yzBcXHxm6nl1fWsDwjQaxulZJgsefFLFTnT+wJF5YdGDzVVXCcXNCK7yGzfaoK5va5lGBA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"4ef1b706082c31a230c920326b9a4980ebfd6a9f0a0f7602cd70c6a904a1c7b7","last_reissued_at":"2026-07-05T08:31:20.886035Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:31:20.886035Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Pruning is Optimal for Learning Sparse Features in High-Dimensions","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"stat.ML","authors_text":"Murat A. Erdogdu, Nuri Mert Vural","submitted_at":"2024-06-12T21:43:12Z","abstract_excerpt":"While it is commonly observed in practice that pruning networks to a certain level of sparsity can improve the quality of the features, a theoretical explanation of this phenomenon remains elusive. In this work, we investigate this by demonstrating that a broad class of statistical models can be optimally learned using pruned neural networks trained with gradient descent, in high-dimensions.\n  We consider learning both single-index and multi-index models of the form $y = \\sigma^*(\\boldsymbol{V}^{\\top} \\boldsymbol{x}) + \\epsilon$, where $\\sigma^*$ is a degree-$p$ polynomial, and $\\boldsymbol{V}"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2406.08658","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/2406.08658/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":"2406.08658","created_at":"2026-07-05T08:31:20.886086+00:00"},{"alias_kind":"arxiv_version","alias_value":"2406.08658v1","created_at":"2026-07-05T08:31:20.886086+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2406.08658","created_at":"2026-07-05T08:31:20.886086+00:00"},{"alias_kind":"pith_short_12","alias_value":"J3Y3OBQIFQY2","created_at":"2026-07-05T08:31:20.886086+00:00"},{"alias_kind":"pith_short_16","alias_value":"J3Y3OBQIFQY2EMGJ","created_at":"2026-07-05T08:31:20.886086+00:00"},{"alias_kind":"pith_short_8","alias_value":"J3Y3OBQI","created_at":"2026-07-05T08:31:20.886086+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2502.09106","citing_title":"Scaling Law for Stochastic Gradient Descent in Quadratically Parameterized Linear Regression","ref_index":44,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/J3Y3OBQIFQY2EMGJEAZGXGSJQD","json":"https://pith.science/pith/J3Y3OBQIFQY2EMGJEAZGXGSJQD.json","graph_json":"https://pith.science/api/pith-number/J3Y3OBQIFQY2EMGJEAZGXGSJQD/graph.json","events_json":"https://pith.science/api/pith-number/J3Y3OBQIFQY2EMGJEAZGXGSJQD/events.json","paper":"https://pith.science/paper/J3Y3OBQI"},"agent_actions":{"view_html":"https://pith.science/pith/J3Y3OBQIFQY2EMGJEAZGXGSJQD","download_json":"https://pith.science/pith/J3Y3OBQIFQY2EMGJEAZGXGSJQD.json","view_paper":"https://pith.science/paper/J3Y3OBQI","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2406.08658&json=true","fetch_graph":"https://pith.science/api/pith-number/J3Y3OBQIFQY2EMGJEAZGXGSJQD/graph.json","fetch_events":"https://pith.science/api/pith-number/J3Y3OBQIFQY2EMGJEAZGXGSJQD/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/J3Y3OBQIFQY2EMGJEAZGXGSJQD/action/timestamp_anchor","attest_storage":"https://pith.science/pith/J3Y3OBQIFQY2EMGJEAZGXGSJQD/action/storage_attestation","attest_author":"https://pith.science/pith/J3Y3OBQIFQY2EMGJEAZGXGSJQD/action/author_attestation","sign_citation":"https://pith.science/pith/J3Y3OBQIFQY2EMGJEAZGXGSJQD/action/citation_signature","submit_replication":"https://pith.science/pith/J3Y3OBQIFQY2EMGJEAZGXGSJQD/action/replication_record"}},"created_at":"2026-07-05T08:31:20.886086+00:00","updated_at":"2026-07-05T08:31:20.886086+00:00"}