{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:RPHRCEJLBY3GAW7NH2G7OJ2XDQ","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"b7e660042d30ada8fb1589a6ead7239b370037ac9f2b5b5f55d84fde3c2260bc","cross_cats_sorted":["cs.LG","math.ST","stat.TH"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"stat.ML","submitted_at":"2026-05-14T17:05:30Z","title_canon_sha256":"45e919745effab06b7d8d5c9d9a01da6c72fc48e270ca2edebcc81d9bce79670"},"schema_version":"1.0","source":{"id":"2605.15082","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.15082","created_at":"2026-05-17T21:18:33Z"},{"alias_kind":"arxiv_version","alias_value":"2605.15082v1","created_at":"2026-05-17T21:18:33Z"},{"alias_kind":"pith_short_12","alias_value":"RPHRCEJLBY3G","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_16","alias_value":"RPHRCEJLBY3GAW7N","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_8","alias_value":"RPHRCEJL","created_at":"2026-05-18T12:33:37Z"}],"graph_snapshots":[{"event_id":"sha256:62fab762d26d31af72ccd0b00688fe67ec6f93d2ad233fb892c2c531ee880f7a","target":"graph","created_at":"2026-05-17T21:57:19Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":4,"items":[{"attestation":"unclaimed","claim_id":"C1","kind":"strongest_claim","source":"verdict.strongest_claim","status":"machine_extracted","text":"the top r-dimensional eigenspace of AGOP provably recovers the central subspace, even in regimes when the prediction error remains large. Specifically, if the target function f* has degree p*, ... subspace recovery occurs in the much lower sample regime n ≍ d^{p+δ} for any δ ∈ (0,1)."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"a low degree p component of f* already carries all relevant directions for prediction, together with unspecified 'reasonable assumptions' on the kernel and the multi-index structure (abstract only)."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"For multi-index polynomials, the top r eigenspace of the AGOP matrix from KRR recovers the central subspace at sample complexity n ~ d^{p+δ} where p is the degree of the informative component."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"The average gradient outer product from kernel ridge regression recovers the central subspace of multi-index models in sample regimes too small for accurate prediction."}],"snapshot_sha256":"45310b90d8bc0c84b1d0a09e96c47bebf99c9878d66db731277a0f6de1693b7e"},"formal_canon":{"evidence_count":2,"snapshot_sha256":"be6fd5681d87fd153a1276402e7e512cbf7be9cc4564aa2c216aaf765ed23e22"},"paper":{"abstract_excerpt":"We study a prototypical situation when a learned predictor can discover useful low-dimensional structure in data, while using fewer samples than are needed for accurate prediction. Specifically, we consider the problem of recovering a multi-index polynomial $f^*(x)=h(Ux)$, with $U\\in\\mathbb{R}^{r\\times d}$ and $r\\ll d$, from finitely many data/label pairs. Importantly, the target function depends on input $x$ only through the projection onto an unknown $r$-dimensional central subspace. The algorithm we analyze is appealingly simple: fit kernel ridge regression (KRR) to the data and compute the","authors_text":"Damek Davis, Dmitriy Drusvyatskiy, Libin Zhu, Maryam Fazel","cross_cats":["cs.LG","math.ST","stat.TH"],"headline":"The average gradient outer product from kernel ridge regression recovers the central subspace of multi-index models in sample regimes too small for accurate prediction.","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"stat.ML","submitted_at":"2026-05-14T17:05:30Z","title":"Average Gradient Outer Product in kernel regression provably recovers the central subspace for multi-index models"},"references":{"count":64,"internal_anchors":0,"resolved_work":64,"sample":[{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":1,"title":"Sgd learning on neural networks: leap complexity and saddle-to-saddle dynamics","work_id":"c710cb1d-e477-4ec1-91b5-a8534000fc4e","year":2023},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":2,"title":"The merged-staircase property: a necessary and nearly sufficient condition for sgd learning of sparse functions on two-layer neural networks","work_id":"4d8fa6c0-826f-4638-986c-6f3c0d96af37","year":2022},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":3,"title":"Classical orthogonal polynomials","work_id":"93ead7bf-64f0-4d79-9e77-66dbc737e138","year":null},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":4,"title":"Springer. 2006, pp. 36–62","work_id":"151a1b61-ae68-4f67-b38b-57c5462d3dea","year":2006},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":5,"title":"Online learning and information exponents: On the importance of batch size, and time/complexity tradeoffs","work_id":"0e969c0f-81e0-4313-af54-714970b4945a","year":2024}],"snapshot_sha256":"d40a3fedb25ffe5dd07b0cccca202bc662cdbfc53b810f036b02bd0af43a4d99"},"source":{"id":"2605.15082","kind":"arxiv","version":1},"verdict":{"created_at":"2026-05-15T03:05:53.439476Z","id":"a08aba0c-35c9-45ae-baca-7e51e7df039e","model_set":{"reader":"grok-4.3"},"one_line_summary":"For multi-index polynomials, the top r eigenspace of the AGOP matrix from KRR recovers the central subspace at sample complexity n ~ d^{p+δ} where p is the degree of the informative component.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"The average gradient outer product from kernel ridge regression recovers the central subspace of multi-index models in sample regimes too small for accurate prediction.","strongest_claim":"the top r-dimensional eigenspace of AGOP provably recovers the central subspace, even in regimes when the prediction error remains large. Specifically, if the target function f* has degree p*, ... subspace recovery occurs in the much lower sample regime n ≍ d^{p+δ} for any δ ∈ (0,1).","weakest_assumption":"a low degree p component of f* already carries all relevant directions for prediction, together with unspecified 'reasonable assumptions' on the kernel and the multi-index structure (abstract only)."}},"verdict_id":"a08aba0c-35c9-45ae-baca-7e51e7df039e"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:7bffd6bd2d86a4640cad7d3bb33069e53d177525a85fffe7acfe3936209672a8","target":"record","created_at":"2026-05-17T21:18:33Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"b7e660042d30ada8fb1589a6ead7239b370037ac9f2b5b5f55d84fde3c2260bc","cross_cats_sorted":["cs.LG","math.ST","stat.TH"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"stat.ML","submitted_at":"2026-05-14T17:05:30Z","title_canon_sha256":"45e919745effab06b7d8d5c9d9a01da6c72fc48e270ca2edebcc81d9bce79670"},"schema_version":"1.0","source":{"id":"2605.15082","kind":"arxiv","version":1}},"canonical_sha256":"8bcf11112b0e36605bed3e8df727571c35cd04e23f6389a3e1f08a728a002ff2","receipt":{"builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"8bcf11112b0e36605bed3e8df727571c35cd04e23f6389a3e1f08a728a002ff2","first_computed_at":"2026-05-17T21:40:25.977506Z","kind":"pith_receipt","last_reissued_at":"2026-05-17T21:57:19.284907Z","receipt_version":"0.2","signature_status":"unsigned_v0"},"source_id":"2605.15082","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:7bffd6bd2d86a4640cad7d3bb33069e53d177525a85fffe7acfe3936209672a8","sha256:62fab762d26d31af72ccd0b00688fe67ec6f93d2ad233fb892c2c531ee880f7a"],"state_sha256":"50208ce57a0e5b9aefbd298ebb08d0ceb90cfac8f01d1f5c71268fa17f68bc37"}