{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:57VM3GD3PIO5IMJ66W476Z3GEZ","short_pith_number":"pith:57VM3GD3","schema_version":"1.0","canonical_sha256":"efeacd987b7a1dd4313ef5b9ff6766267af3f03cc4e0702570c2ef8016267ca2","source":{"kind":"arxiv","id":"2411.08798","version":2},"attestation_state":"computed","paper":{"title":"Learning Gaussian Multi-Index Models with Gradient Flow: Time Complexity and Directional Convergence","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.NE","math.AG"],"primary_cat":"cs.LG","authors_text":"Amire Bendjeddou, Berfin \\c{S}im\\c{s}ek, Daniel Hsu","submitted_at":"2024-11-13T17:25:25Z","abstract_excerpt":"This work focuses on the gradient flow dynamics of a neural network model that uses correlation loss to approximate a multi-index function on high-dimensional standard Gaussian data. Specifically, the multi-index function we consider is a sum of neurons $f^*(x) \\!=\\! \\sum_{j=1}^k \\! \\sigma^*(v_j^T x)$ where $v_1, \\dots, v_k$ are unit vectors, and $\\sigma^*$ lacks the first and second Hermite polynomials in its Hermite expansion. It is known that, for the single-index case ($k\\!=\\!1$), overcoming the search phase requires polynomial time complexity. We first generalize this result to multi-inde"},"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":"2411.08798","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-11-13T17:25:25Z","cross_cats_sorted":["cs.NE","math.AG"],"title_canon_sha256":"b2422a18c7a40e919744a2ad728ab82dac7d99fa69549b09f9109a4c55cd440a","abstract_canon_sha256":"f520dd461ce59495fc2ebc417ef070736104878e2ac89ec3b5d9a6689b74e77a"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:28:26.209289Z","signature_b64":"JluWZaCeOYCrvtwyNnClBfdenGcooCxy5k/foylo8USUewYOy9oHQDCoOlxd3aZx+lkNw5tqqo59AGVqEkEICQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"efeacd987b7a1dd4313ef5b9ff6766267af3f03cc4e0702570c2ef8016267ca2","last_reissued_at":"2026-07-05T10:28:26.208733Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:28:26.208733Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Learning Gaussian Multi-Index Models with Gradient Flow: Time Complexity and Directional Convergence","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.NE","math.AG"],"primary_cat":"cs.LG","authors_text":"Amire Bendjeddou, Berfin \\c{S}im\\c{s}ek, Daniel Hsu","submitted_at":"2024-11-13T17:25:25Z","abstract_excerpt":"This work focuses on the gradient flow dynamics of a neural network model that uses correlation loss to approximate a multi-index function on high-dimensional standard Gaussian data. Specifically, the multi-index function we consider is a sum of neurons $f^*(x) \\!=\\! \\sum_{j=1}^k \\! \\sigma^*(v_j^T x)$ where $v_1, \\dots, v_k$ are unit vectors, and $\\sigma^*$ lacks the first and second Hermite polynomials in its Hermite expansion. It is known that, for the single-index case ($k\\!=\\!1$), overcoming the search phase requires polynomial time complexity. We first generalize this result to multi-inde"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2411.08798","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/2411.08798/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":"2411.08798","created_at":"2026-07-05T10:28:26.208789+00:00"},{"alias_kind":"arxiv_version","alias_value":"2411.08798v2","created_at":"2026-07-05T10:28:26.208789+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2411.08798","created_at":"2026-07-05T10:28:26.208789+00:00"},{"alias_kind":"pith_short_12","alias_value":"57VM3GD3PIO5","created_at":"2026-07-05T10:28:26.208789+00:00"},{"alias_kind":"pith_short_16","alias_value":"57VM3GD3PIO5IMJ6","created_at":"2026-07-05T10:28:26.208789+00:00"},{"alias_kind":"pith_short_8","alias_value":"57VM3GD3","created_at":"2026-07-05T10:28:26.208789+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2506.14951","citing_title":"Flat Channels to Infinity in Neural Loss Landscapes","ref_index":22,"is_internal_anchor":false},{"citing_arxiv_id":"2605.14567","citing_title":"Scaling Laws from Sequential Feature Recovery: A Solvable Hierarchical Model","ref_index":94,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/57VM3GD3PIO5IMJ66W476Z3GEZ","json":"https://pith.science/pith/57VM3GD3PIO5IMJ66W476Z3GEZ.json","graph_json":"https://pith.science/api/pith-number/57VM3GD3PIO5IMJ66W476Z3GEZ/graph.json","events_json":"https://pith.science/api/pith-number/57VM3GD3PIO5IMJ66W476Z3GEZ/events.json","paper":"https://pith.science/paper/57VM3GD3"},"agent_actions":{"view_html":"https://pith.science/pith/57VM3GD3PIO5IMJ66W476Z3GEZ","download_json":"https://pith.science/pith/57VM3GD3PIO5IMJ66W476Z3GEZ.json","view_paper":"https://pith.science/paper/57VM3GD3","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2411.08798&json=true","fetch_graph":"https://pith.science/api/pith-number/57VM3GD3PIO5IMJ66W476Z3GEZ/graph.json","fetch_events":"https://pith.science/api/pith-number/57VM3GD3PIO5IMJ66W476Z3GEZ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/57VM3GD3PIO5IMJ66W476Z3GEZ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/57VM3GD3PIO5IMJ66W476Z3GEZ/action/storage_attestation","attest_author":"https://pith.science/pith/57VM3GD3PIO5IMJ66W476Z3GEZ/action/author_attestation","sign_citation":"https://pith.science/pith/57VM3GD3PIO5IMJ66W476Z3GEZ/action/citation_signature","submit_replication":"https://pith.science/pith/57VM3GD3PIO5IMJ66W476Z3GEZ/action/replication_record"}},"created_at":"2026-07-05T10:28:26.208789+00:00","updated_at":"2026-07-05T10:28:26.208789+00:00"}