{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:OIPHSS7TJ6NNB6YXGDEQE3Q5KZ","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":"4f5e35460f433f2770a89e63256e4742a2edfbeeb4ba2ed24fd0929891281947","cross_cats_sorted":["cs.LG","math.PR","math.ST","stat.TH"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"stat.ML","submitted_at":"2026-05-14T08:37:28Z","title_canon_sha256":"ec3bf8972b912f87282da59ffb84651525cf10c2cfbddf538ad7584caffec381"},"schema_version":"1.0","source":{"id":"2605.14567","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.14567","created_at":"2026-05-17T23:39:05Z"},{"alias_kind":"arxiv_version","alias_value":"2605.14567v1","created_at":"2026-05-17T23:39:05Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.14567","created_at":"2026-05-17T23:39:05Z"},{"alias_kind":"pith_short_12","alias_value":"OIPHSS7TJ6NN","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_16","alias_value":"OIPHSS7TJ6NNB6YX","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_8","alias_value":"OIPHSS7T","created_at":"2026-05-18T12:33:37Z"}],"graph_snapshots":[{"event_id":"sha256:316109213081cf0e0a7a12a6ca872f90f970a8c5e9c4ed10b1c703cda328a86e","target":"graph","created_at":"2026-05-17T23:39:05Z","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":"aggregating these transitions yields an explicit power-law decay of the prediction error"},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"the high-dimensional target admits a representation as a combination of latent compositional features whose weights decrease as a power law, and that the layer-wise spectral algorithm is specifically adapted to this compositional structure"},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"A solvable hierarchical model with power-law feature strengths yields explicit power-law scaling of prediction error through sequential recovery of latent directions by a layer-wise spectral algorithm."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"A layer-wise spectral algorithm on a hierarchical target with power-law feature weights recovers latent directions sequentially and aggregates their sharp thresholds into an explicit power-law decay of prediction error."}],"snapshot_sha256":"f80eb6c129438bd9d298e5e3fd1390cce232e3daf334947995d4fcfcfa3afbc3"},"formal_canon":{"evidence_count":2,"snapshot_sha256":"4e33ba0197c01b1b0f2800e936ec3223b7a49c6d00f2f8331348b6c633fbca5f"},"paper":{"abstract_excerpt":"We propose a simple mechanism by which scaling laws emerge from feature learning in multi-layer networks. We study a high-dimensional hierarchical target that is a globally high-degree function, but that can be represented by a combination of latent compositional features whose weights decrease as a power law. We show that a layer-wise spectral algorithm adapted to this compositional structure achieves improved scaling relative to shallow, non-adaptive methods, and recovers the latent directions sequentially: strong features become detectable at small sample sizes, while weaker features requir","authors_text":"Arie Wortsman-Zurich, Bruno Loureiro, Florent Krzakala, Hugo Tabanelli, Yatin Dandi","cross_cats":["cs.LG","math.PR","math.ST","stat.TH"],"headline":"A layer-wise spectral algorithm on a hierarchical target with power-law feature weights recovers latent directions sequentially and aggregates their sharp thresholds into an explicit power-law decay of prediction error.","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"stat.ML","submitted_at":"2026-05-14T08:37:28Z","title":"Scaling Laws from Sequential Feature Recovery: A Solvable Hierarchical Model"},"references":{"count":187,"internal_anchors":7,"resolved_work":187,"sample":[{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":1,"title":"Random matrix methods for machine learning , author=. 2022 , publisher=","work_id":"7cb35935-4ffa-43b3-b4ff-744e845a45b5","year":2022},{"cited_arxiv_id":"1011.3027","doi":"","is_internal_anchor":true,"ref_index":2,"title":"Introduction to the non-asymptotic analysis of random matrices","work_id":"695fe191-d99a-4805-9596-420902e88b7f","year":null},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":3,"title":"Applied and Computational Harmonic Analysis , volume=","work_id":"e917bb1e-a824-4914-a667-682624bbbf0b","year":2022},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":4,"title":"Bert: Pre-training of deep bidirectional transformers for language understanding , author=. Proceedings of the 2019 conference of the North American chapter of the association for computational lingui","work_id":"8d56cd15-b543-4ffe-84e6-e4e3e88c7232","year":2019},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":5,"title":"arXiv preprint arXiv:2305.15501 , year=","work_id":"a55b9778-aa9e-4c05-a613-6263dc711e0e","year":null}],"snapshot_sha256":"5f77b15a5e23c27119ea3f35ba551474b1d175904c1432f970846c605213a3c6"},"source":{"id":"2605.14567","kind":"arxiv","version":1},"verdict":{"created_at":"2026-05-15T01:35:40.345991Z","id":"6e1da657-0718-4853-a364-30dbde508cf1","model_set":{"reader":"grok-4.3"},"one_line_summary":"A solvable hierarchical model with power-law feature strengths yields explicit power-law scaling of prediction error through sequential recovery of latent directions by a layer-wise spectral algorithm.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"A layer-wise spectral algorithm on a hierarchical target with power-law feature weights recovers latent directions sequentially and aggregates their sharp thresholds into an explicit power-law decay of prediction error.","strongest_claim":"aggregating these transitions yields an explicit power-law decay of the prediction error","weakest_assumption":"the high-dimensional target admits a representation as a combination of latent compositional features whose weights decrease as a power law, and that the layer-wise spectral algorithm is specifically adapted to this compositional structure"}},"verdict_id":"6e1da657-0718-4853-a364-30dbde508cf1"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:debb7699f844dc6f39c08f4c9bd7b8efb0f3425a12d9b57f22b9873d54ab4d92","target":"record","created_at":"2026-05-17T23:39:05Z","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":"4f5e35460f433f2770a89e63256e4742a2edfbeeb4ba2ed24fd0929891281947","cross_cats_sorted":["cs.LG","math.PR","math.ST","stat.TH"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"stat.ML","submitted_at":"2026-05-14T08:37:28Z","title_canon_sha256":"ec3bf8972b912f87282da59ffb84651525cf10c2cfbddf538ad7584caffec381"},"schema_version":"1.0","source":{"id":"2605.14567","kind":"arxiv","version":1}},"canonical_sha256":"721e794bf34f9ad0fb1730c9026e1d56748be3ba441a947fe8f544ba55b7f10a","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"721e794bf34f9ad0fb1730c9026e1d56748be3ba441a947fe8f544ba55b7f10a","first_computed_at":"2026-05-17T23:39:05.519254Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-17T23:39:05.519254Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"uda/BKS9OUh1FSY0QqYMtIFUh5G1Kjvb2Xl2TPSPCKtP+kcxMJTK1gq0lUmZKWrPPyZw93Bv+lnGmq5W6apfAg==","signature_status":"signed_v1","signed_at":"2026-05-17T23:39:05.519997Z","signed_message":"canonical_sha256_bytes"},"source_id":"2605.14567","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:debb7699f844dc6f39c08f4c9bd7b8efb0f3425a12d9b57f22b9873d54ab4d92","sha256:316109213081cf0e0a7a12a6ca872f90f970a8c5e9c4ed10b1c703cda328a86e"],"state_sha256":"7adfb45399be387322de24e91ce10b5783d0eb2247dfd3bd49411fde324c7a8c"}