{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:DMMM25JRWHB5BF6HFZCCG6P3H6","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":"192b42be698d2b273a2625dff2f2e745a58fa82ae6084e9fbe9494fa15d76360","cross_cats_sorted":["cond-mat.dis-nn","cond-mat.stat-mech","cs.LG","math.PR"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"stat.ML","submitted_at":"2026-05-16T10:03:41Z","title_canon_sha256":"43eea5453ba470f353dc6b1cc1aae26ab974026a57d03cabe76876b04b100201"},"schema_version":"1.0","source":{"id":"2605.16913","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.16913","created_at":"2026-05-20T00:03:29Z"},{"alias_kind":"arxiv_version","alias_value":"2605.16913v1","created_at":"2026-05-20T00:03:29Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.16913","created_at":"2026-05-20T00:03:29Z"},{"alias_kind":"pith_short_12","alias_value":"DMMM25JRWHB5","created_at":"2026-05-20T00:03:29Z"},{"alias_kind":"pith_short_16","alias_value":"DMMM25JRWHB5BF6H","created_at":"2026-05-20T00:03:29Z"},{"alias_kind":"pith_short_8","alias_value":"DMMM25JR","created_at":"2026-05-20T00:03:29Z"}],"graph_snapshots":[{"event_id":"sha256:d1e423ef886e31680c9531821f8a3c6b2e48046c2a9c9fcf794079e5180c15bf","target":"graph","created_at":"2026-05-20T00:03:29Z","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":"For isotropic and high-dimensional inputs, classification based on phase information alone is a genuinely hard task: online SGD cannot distinguish the structured inputs from noise within n ≪ N³ steps, but needs at least n ≫ N³ log²N steps."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"The introduced synthetic data model for translation-invariant inputs faithfully captures the interaction between amplitudes, phases, and learning dynamics without introducing artifacts that invalidate the hardness result or the power-law acceleration claim."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"Neural networks prioritize amplitude over phase in Fourier space during training on translation-invariant data; power-law spectra accelerate phase learning despite not aiding classification."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"Online SGD cannot learn phase-only classification on isotropic high-dimensional inputs before order N cubed steps, but power-law spectra accelerate it substantially."}],"snapshot_sha256":"6611edc52f38979eec3582f9a7b0ab57961b5a561a4cdd732d35d2bb35bad8ad"},"formal_canon":{"evidence_count":2,"snapshot_sha256":"15b45c0048917a93b5b8d40ca6a64ac866edba30c2d610c80219203f27e1ad76"},"integrity":{"available":true,"clean":true,"detectors_run":[{"findings_count":0,"name":"cited_work_retraction","ran_at":"2026-05-19T20:52:02.966853Z","status":"completed","version":"1.0.0"},{"findings_count":0,"name":"doi_title_agreement","ran_at":"2026-05-19T20:01:18.966218Z","status":"completed","version":"1.0.0"},{"findings_count":0,"name":"doi_compliance","ran_at":"2026-05-19T19:50:42.466159Z","status":"completed","version":"1.0.0"},{"findings_count":0,"name":"claim_evidence","ran_at":"2026-05-19T18:41:56.267720Z","status":"completed","version":"1.0.0"},{"findings_count":0,"name":"ai_meta_artifact","ran_at":"2026-05-19T18:33:26.347687Z","status":"skipped","version":"1.0.0"}],"endpoint":"/pith/2605.16913/integrity.json","findings":[],"snapshot_sha256":"bad1206fe8f822f674b123f4c5e618faf1a708c6a46cb649529c6270280635ff","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Neural networks trained with gradient-based methods exhibit a strong simplicity bias: they learn simpler statistical features of their data before moving to more complex features. Previous analyses of this phenomenon have largely focused on settings with (quasi-)isotropic inputs. In this work, we study the simplicity bias from a Fourier perspective, which allows us to include two key features of natural images in the analysis: approximate translation-invariance and power-law spectra. We first show experimentally that simple neural networks trained on image classification tasks first rely on am","authors_text":"Claudia Merger, Fabiola Ricci, Sebastian Goldt","cross_cats":["cond-mat.dis-nn","cond-mat.stat-mech","cs.LG","math.PR"],"headline":"Online SGD cannot learn phase-only classification on isotropic high-dimensional inputs before order N cubed steps, but power-law spectra accelerate it substantially.","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"stat.ML","submitted_at":"2026-05-16T10:03:41Z","title":"A Fourier perspective on the learning dynamics of neural networks: from sample complexities to mechanistic insights"},"references":{"count":69,"internal_anchors":0,"resolved_work":69,"sample":[{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":1,"title":"et al.SGD on Neural Networks Learns Functions of Increasing ComplexityinAdvances in Neural Information Processing Systems32(2019), 3491–3501","work_id":"53683ea2-598d-4167-a3f8-cd1ee846db79","year":2019},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":2,"title":"Ingrosso, A. & Goldt, S. Data-driven emergence of convolutional structure in neural networks. Proceedings of the National Academy of Sciences119(2022)","work_id":"17bb55d4-69d2-4baa-ae29-e65e92b2d05f","year":2022},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":3,"title":"& Goldt, S.Neural networks trained with SGD learn distributions of increasing complexityinInternational Conference on Machine Learning(2023), 28843–28863","work_id":"ef98ae7f-c1c1-4dc6-a96c-003eb64b3d6c","year":2023},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":4,"title":"& Goldt, S.A distributional simplicity bias in the learning dynamics of transformersinAdvances in Neural Information Processing Systems37(2024), 96207–96228","work_id":"f1b76ea0-2af5-45de-9547-0e463c9ccdea","year":2024},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":5,"title":"& Fern, X.Neural Networks Learn Statistics of Increasing Complexityin (arXiv, 2024)","work_id":"3d70ac82-d6a4-4a01-b1ac-fa87c1aa4d57","year":2024}],"snapshot_sha256":"84229301d22c2d48eda220452f8ebec61a76b86afad4ad887306e0c207f93bf3"},"source":{"id":"2605.16913","kind":"arxiv","version":1},"verdict":{"created_at":"2026-05-19T19:36:43.567585Z","id":"b07deb0d-611c-4111-bf7f-005316dcd5e4","model_set":{"reader":"grok-4.3"},"one_line_summary":"Neural networks prioritize amplitude over phase in Fourier space during training on translation-invariant data; power-law spectra accelerate phase learning despite not aiding classification.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"Online SGD cannot learn phase-only classification on isotropic high-dimensional inputs before order N cubed steps, but power-law spectra accelerate it substantially.","strongest_claim":"For isotropic and high-dimensional inputs, classification based on phase information alone is a genuinely hard task: online SGD cannot distinguish the structured inputs from noise within n ≪ N³ steps, but needs at least n ≫ N³ log²N steps.","weakest_assumption":"The introduced synthetic data model for translation-invariant inputs faithfully captures the interaction between amplitudes, phases, and learning dynamics without introducing artifacts that invalidate the hardness result or the power-law acceleration claim."}},"verdict_id":"b07deb0d-611c-4111-bf7f-005316dcd5e4"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:61c1621b681a9de64de1e6c7f160ebd2932b54f39c580fd5349616ffbe3e86fc","target":"record","created_at":"2026-05-20T00:03:29Z","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":"192b42be698d2b273a2625dff2f2e745a58fa82ae6084e9fbe9494fa15d76360","cross_cats_sorted":["cond-mat.dis-nn","cond-mat.stat-mech","cs.LG","math.PR"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"stat.ML","submitted_at":"2026-05-16T10:03:41Z","title_canon_sha256":"43eea5453ba470f353dc6b1cc1aae26ab974026a57d03cabe76876b04b100201"},"schema_version":"1.0","source":{"id":"2605.16913","kind":"arxiv","version":1}},"canonical_sha256":"1b18cd7531b1c3d097c72e442379fb3facb2fa2df17c24134cd20b8db0c6f24c","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"1b18cd7531b1c3d097c72e442379fb3facb2fa2df17c24134cd20b8db0c6f24c","first_computed_at":"2026-05-20T00:03:29.862957Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-20T00:03:29.862957Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"FwFnFSo0EW7z2nU2ZxTY37H8duGjafLG6FNIsGoV+Ncz/limsM9gU9f8M1nCdI1PeYlUaoQ24gCTbIzsR4iAAw==","signature_status":"signed_v1","signed_at":"2026-05-20T00:03:29.863874Z","signed_message":"canonical_sha256_bytes"},"source_id":"2605.16913","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:61c1621b681a9de64de1e6c7f160ebd2932b54f39c580fd5349616ffbe3e86fc","sha256:d1e423ef886e31680c9531821f8a3c6b2e48046c2a9c9fcf794079e5180c15bf"],"state_sha256":"71a18fa6c59ebe1cd61a700851b80078e1534fffb6b1f19161fc907eacd3a82e"}