{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2017:MVHS7BIG6LW4KL6JDVQG67MYG3","short_pith_number":"pith:MVHS7BIG","schema_version":"1.0","canonical_sha256":"654f2f8506f2edc52fc91d606f7d9836e67ccb29c162b84a5de77ee124dc0a7b","source":{"kind":"arxiv","id":"1712.09203","version":5},"attestation_state":"computed","paper":{"title":"Algorithmic Regularization in Over-parameterized Matrix Sensing and Neural Networks with Quadratic Activations","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.DS","math.OC","stat.ML"],"primary_cat":"cs.LG","authors_text":"Hongyang Zhang, Tengyu Ma, Yuanzhi Li","submitted_at":"2017-12-26T08:04:43Z","abstract_excerpt":"We show that the gradient descent algorithm provides an implicit regularization effect in the learning of over-parameterized matrix factorization models and one-hidden-layer neural networks with quadratic activations. Concretely, we show that given $\\tilde{O}(dr^{2})$ random linear measurements of a rank $r$ positive semidefinite matrix $X^{\\star}$, we can recover $X^{\\star}$ by parameterizing it by $UU^\\top$ with $U\\in \\mathbb R^{d\\times d}$ and minimizing the squared loss, even if $r \\ll d$. We prove that starting from a small initialization, gradient descent recovers $X^{\\star}$ in $\\tilde{"},"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":"1712.09203","kind":"arxiv","version":5},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2017-12-26T08:04:43Z","cross_cats_sorted":["cs.DS","math.OC","stat.ML"],"title_canon_sha256":"d14addae24dc3a7a554ea2e6add3264a7bdec192ee8c087c47e5f0c13c357886","abstract_canon_sha256":"d2c7e950feb19efe61230fd4a01b8f8773ed1751f6a05292373da92e559fc9a4"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-17T23:54:03.141861Z","signature_b64":"1AnLbLdMFibE5DTQy0x1L2QJ970WQeAF1NNvsmIRy90kVErivyOfFUdf3gmBlsfAvNBYlrXZVOJ3twbA3Xb1AA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"654f2f8506f2edc52fc91d606f7d9836e67ccb29c162b84a5de77ee124dc0a7b","last_reissued_at":"2026-05-17T23:54:03.141245Z","signature_status":"signed_v1","first_computed_at":"2026-05-17T23:54:03.141245Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Algorithmic Regularization in Over-parameterized Matrix Sensing and Neural Networks with Quadratic Activations","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.DS","math.OC","stat.ML"],"primary_cat":"cs.LG","authors_text":"Hongyang Zhang, Tengyu Ma, Yuanzhi Li","submitted_at":"2017-12-26T08:04:43Z","abstract_excerpt":"We show that the gradient descent algorithm provides an implicit regularization effect in the learning of over-parameterized matrix factorization models and one-hidden-layer neural networks with quadratic activations. Concretely, we show that given $\\tilde{O}(dr^{2})$ random linear measurements of a rank $r$ positive semidefinite matrix $X^{\\star}$, we can recover $X^{\\star}$ by parameterizing it by $UU^\\top$ with $U\\in \\mathbb R^{d\\times d}$ and minimizing the squared loss, even if $r \\ll d$. We prove that starting from a small initialization, gradient descent recovers $X^{\\star}$ in $\\tilde{"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1712.09203","kind":"arxiv","version":5},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"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":"1712.09203","created_at":"2026-05-17T23:54:03.141358+00:00"},{"alias_kind":"arxiv_version","alias_value":"1712.09203v5","created_at":"2026-05-17T23:54:03.141358+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1712.09203","created_at":"2026-05-17T23:54:03.141358+00:00"},{"alias_kind":"pith_short_12","alias_value":"MVHS7BIG6LW4","created_at":"2026-05-18T12:31:31.346846+00:00"},{"alias_kind":"pith_short_16","alias_value":"MVHS7BIG6LW4KL6J","created_at":"2026-05-18T12:31:31.346846+00:00"},{"alias_kind":"pith_short_8","alias_value":"MVHS7BIG","created_at":"2026-05-18T12:31:31.346846+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/MVHS7BIG6LW4KL6JDVQG67MYG3","json":"https://pith.science/pith/MVHS7BIG6LW4KL6JDVQG67MYG3.json","graph_json":"https://pith.science/api/pith-number/MVHS7BIG6LW4KL6JDVQG67MYG3/graph.json","events_json":"https://pith.science/api/pith-number/MVHS7BIG6LW4KL6JDVQG67MYG3/events.json","paper":"https://pith.science/paper/MVHS7BIG"},"agent_actions":{"view_html":"https://pith.science/pith/MVHS7BIG6LW4KL6JDVQG67MYG3","download_json":"https://pith.science/pith/MVHS7BIG6LW4KL6JDVQG67MYG3.json","view_paper":"https://pith.science/paper/MVHS7BIG","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=1712.09203&json=true","fetch_graph":"https://pith.science/api/pith-number/MVHS7BIG6LW4KL6JDVQG67MYG3/graph.json","fetch_events":"https://pith.science/api/pith-number/MVHS7BIG6LW4KL6JDVQG67MYG3/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/MVHS7BIG6LW4KL6JDVQG67MYG3/action/timestamp_anchor","attest_storage":"https://pith.science/pith/MVHS7BIG6LW4KL6JDVQG67MYG3/action/storage_attestation","attest_author":"https://pith.science/pith/MVHS7BIG6LW4KL6JDVQG67MYG3/action/author_attestation","sign_citation":"https://pith.science/pith/MVHS7BIG6LW4KL6JDVQG67MYG3/action/citation_signature","submit_replication":"https://pith.science/pith/MVHS7BIG6LW4KL6JDVQG67MYG3/action/replication_record"}},"created_at":"2026-05-17T23:54:03.141358+00:00","updated_at":"2026-05-17T23:54:03.141358+00:00"}