{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:6LL54QOTFZD7UZXJCRDYSIOMIF","short_pith_number":"pith:6LL54QOT","schema_version":"1.0","canonical_sha256":"f2d7de41d32e47fa66e914478921cc415385199f8b8febef0114ea7f09ce8d37","source":{"kind":"arxiv","id":"2410.04642","version":3},"attestation_state":"computed","paper":{"title":"The Optimization Landscape of SGD Across the Feature Learning Strength","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["stat.ML"],"primary_cat":"cs.LG","authors_text":"Alexander Atanasov, Alexandru Meterez, Cengiz Pehlevan, James B. Simon","submitted_at":"2024-10-06T22:30:14Z","abstract_excerpt":"We consider neural networks (NNs) where the final layer is down-scaled by a fixed hyperparameter $\\gamma$. Recent work has identified $\\gamma$ as controlling the strength of feature learning. As $\\gamma$ increases, network evolution changes from \"lazy\" kernel dynamics to \"rich\" feature-learning dynamics, with a host of associated benefits including improved performance on common tasks. In this work, we conduct a thorough empirical investigation of the effect of scaling $\\gamma$ across a variety of models and datasets in the online training setting. We first examine the interaction of $\\gamma$ "},"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":"2410.04642","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-10-06T22:30:14Z","cross_cats_sorted":["stat.ML"],"title_canon_sha256":"8491fe178cca680313721c656021c913418b317c550706fc72a25cfe51c4739c","abstract_canon_sha256":"dafa59d4b34672a25d0c061139c35936536dff7286ee78a51a19dec62cb9352c"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:22:06.368574Z","signature_b64":"LcYtM+q4GWnVJ0qdBQP20N06awwrNVYrn4iQkB41X9czcYMpwySUFYYzGD3EAjdiYGX2U1mjrTMX9nNisoHOAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"f2d7de41d32e47fa66e914478921cc415385199f8b8febef0114ea7f09ce8d37","last_reissued_at":"2026-07-05T10:22:06.368020Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:22:06.368020Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"The Optimization Landscape of SGD Across the Feature Learning Strength","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["stat.ML"],"primary_cat":"cs.LG","authors_text":"Alexander Atanasov, Alexandru Meterez, Cengiz Pehlevan, James B. Simon","submitted_at":"2024-10-06T22:30:14Z","abstract_excerpt":"We consider neural networks (NNs) where the final layer is down-scaled by a fixed hyperparameter $\\gamma$. Recent work has identified $\\gamma$ as controlling the strength of feature learning. As $\\gamma$ increases, network evolution changes from \"lazy\" kernel dynamics to \"rich\" feature-learning dynamics, with a host of associated benefits including improved performance on common tasks. In this work, we conduct a thorough empirical investigation of the effect of scaling $\\gamma$ across a variety of models and datasets in the online training setting. We first examine the interaction of $\\gamma$ "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2410.04642","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/2410.04642/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":"2410.04642","created_at":"2026-07-05T10:22:06.368084+00:00"},{"alias_kind":"arxiv_version","alias_value":"2410.04642v3","created_at":"2026-07-05T10:22:06.368084+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2410.04642","created_at":"2026-07-05T10:22:06.368084+00:00"},{"alias_kind":"pith_short_12","alias_value":"6LL54QOTFZD7","created_at":"2026-07-05T10:22:06.368084+00:00"},{"alias_kind":"pith_short_16","alias_value":"6LL54QOTFZD7UZXJ","created_at":"2026-07-05T10:22:06.368084+00:00"},{"alias_kind":"pith_short_8","alias_value":"6LL54QOT","created_at":"2026-07-05T10:22:06.368084+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":3,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.01288","citing_title":"A Theory of Saddle Escape in Deep Nonlinear Networks","ref_index":9,"is_internal_anchor":false},{"citing_arxiv_id":"2605.01288","citing_title":"A Theory of Saddle Escape in Deep Nonlinear Networks","ref_index":9,"is_internal_anchor":false},{"citing_arxiv_id":"2605.01288","citing_title":"A Theory of Saddle Escape in Deep Nonlinear Networks","ref_index":9,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/6LL54QOTFZD7UZXJCRDYSIOMIF","json":"https://pith.science/pith/6LL54QOTFZD7UZXJCRDYSIOMIF.json","graph_json":"https://pith.science/api/pith-number/6LL54QOTFZD7UZXJCRDYSIOMIF/graph.json","events_json":"https://pith.science/api/pith-number/6LL54QOTFZD7UZXJCRDYSIOMIF/events.json","paper":"https://pith.science/paper/6LL54QOT"},"agent_actions":{"view_html":"https://pith.science/pith/6LL54QOTFZD7UZXJCRDYSIOMIF","download_json":"https://pith.science/pith/6LL54QOTFZD7UZXJCRDYSIOMIF.json","view_paper":"https://pith.science/paper/6LL54QOT","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2410.04642&json=true","fetch_graph":"https://pith.science/api/pith-number/6LL54QOTFZD7UZXJCRDYSIOMIF/graph.json","fetch_events":"https://pith.science/api/pith-number/6LL54QOTFZD7UZXJCRDYSIOMIF/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/6LL54QOTFZD7UZXJCRDYSIOMIF/action/timestamp_anchor","attest_storage":"https://pith.science/pith/6LL54QOTFZD7UZXJCRDYSIOMIF/action/storage_attestation","attest_author":"https://pith.science/pith/6LL54QOTFZD7UZXJCRDYSIOMIF/action/author_attestation","sign_citation":"https://pith.science/pith/6LL54QOTFZD7UZXJCRDYSIOMIF/action/citation_signature","submit_replication":"https://pith.science/pith/6LL54QOTFZD7UZXJCRDYSIOMIF/action/replication_record"}},"created_at":"2026-07-05T10:22:06.368084+00:00","updated_at":"2026-07-05T10:22:06.368084+00:00"}