{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:QBDFITGWUHWNQKHLVMA3TCK7PT","short_pith_number":"pith:QBDFITGW","schema_version":"1.0","canonical_sha256":"8046544cd6a1ecd828ebab01b9895f7cdd3b673e36cbed318047b6ddb74e08ae","source":{"kind":"arxiv","id":"2606.02993","version":1},"attestation_state":"computed","paper":{"title":"Neural Networks Provably Learn Spectral Representations for Group Composition","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["math.OC","math.RT","math.ST","stat.ML","stat.TH"],"primary_cat":"cs.LG","authors_text":"Fengzhuo Zhang, Jianliang He, Leda Wang, Siyu Chen, Zhuoran Yang","submitted_at":"2026-06-02T01:04:21Z","abstract_excerpt":"Understanding how structured internal structure emerges during neural network training is central to the study of deep learning. We investigate this phenomenon through the group composition task, where a two-layer neural network is trained to predict $g_1 \\star g_2$ for elements of a finite group $G$. By lifting the projected gradient flow to the Fourier domain, we demonstrate that the training dynamics are governed by a Riemannian gradient ascent on a representation-theoretic energy functional. We prove that, under random initialization, this flow drives each neuron to converge almost surely "},"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":"2606.02993","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-06-02T01:04:21Z","cross_cats_sorted":["math.OC","math.RT","math.ST","stat.ML","stat.TH"],"title_canon_sha256":"a44a6edaccb5a456c1e3f7e6a3363f6993c617a8b719dedfebecc37255567aef","abstract_canon_sha256":"e233cd0dcaa1e23fc62f5d6f8b936284ef879ffd5c646f701fe9275bbfeeaef0"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-03T01:05:28.757341Z","signature_b64":"Gxi190UJgj/dIY3KZibwl8Ig5rkq3TdMrbJeZPNDoJugHMW7Vwz/jh8V8Zc79r2ZLv0LM9UlrziT0ofALVyRCg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"8046544cd6a1ecd828ebab01b9895f7cdd3b673e36cbed318047b6ddb74e08ae","last_reissued_at":"2026-06-03T01:05:28.756924Z","signature_status":"signed_v1","first_computed_at":"2026-06-03T01:05:28.756924Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Neural Networks Provably Learn Spectral Representations for Group Composition","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["math.OC","math.RT","math.ST","stat.ML","stat.TH"],"primary_cat":"cs.LG","authors_text":"Fengzhuo Zhang, Jianliang He, Leda Wang, Siyu Chen, Zhuoran Yang","submitted_at":"2026-06-02T01:04:21Z","abstract_excerpt":"Understanding how structured internal structure emerges during neural network training is central to the study of deep learning. We investigate this phenomenon through the group composition task, where a two-layer neural network is trained to predict $g_1 \\star g_2$ for elements of a finite group $G$. By lifting the projected gradient flow to the Fourier domain, we demonstrate that the training dynamics are governed by a Riemannian gradient ascent on a representation-theoretic energy functional. We prove that, under random initialization, this flow drives each neuron to converge almost surely "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.02993","kind":"arxiv","version":1},"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/2606.02993/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":"2606.02993","created_at":"2026-06-03T01:05:28.756982+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.02993v1","created_at":"2026-06-03T01:05:28.756982+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.02993","created_at":"2026-06-03T01:05:28.756982+00:00"},{"alias_kind":"pith_short_12","alias_value":"QBDFITGWUHWN","created_at":"2026-06-03T01:05:28.756982+00:00"},{"alias_kind":"pith_short_16","alias_value":"QBDFITGWUHWNQKHL","created_at":"2026-06-03T01:05:28.756982+00:00"},{"alias_kind":"pith_short_8","alias_value":"QBDFITGW","created_at":"2026-06-03T01:05:28.756982+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/QBDFITGWUHWNQKHLVMA3TCK7PT","json":"https://pith.science/pith/QBDFITGWUHWNQKHLVMA3TCK7PT.json","graph_json":"https://pith.science/api/pith-number/QBDFITGWUHWNQKHLVMA3TCK7PT/graph.json","events_json":"https://pith.science/api/pith-number/QBDFITGWUHWNQKHLVMA3TCK7PT/events.json","paper":"https://pith.science/paper/QBDFITGW"},"agent_actions":{"view_html":"https://pith.science/pith/QBDFITGWUHWNQKHLVMA3TCK7PT","download_json":"https://pith.science/pith/QBDFITGWUHWNQKHLVMA3TCK7PT.json","view_paper":"https://pith.science/paper/QBDFITGW","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.02993&json=true","fetch_graph":"https://pith.science/api/pith-number/QBDFITGWUHWNQKHLVMA3TCK7PT/graph.json","fetch_events":"https://pith.science/api/pith-number/QBDFITGWUHWNQKHLVMA3TCK7PT/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/QBDFITGWUHWNQKHLVMA3TCK7PT/action/timestamp_anchor","attest_storage":"https://pith.science/pith/QBDFITGWUHWNQKHLVMA3TCK7PT/action/storage_attestation","attest_author":"https://pith.science/pith/QBDFITGWUHWNQKHLVMA3TCK7PT/action/author_attestation","sign_citation":"https://pith.science/pith/QBDFITGWUHWNQKHLVMA3TCK7PT/action/citation_signature","submit_replication":"https://pith.science/pith/QBDFITGWUHWNQKHLVMA3TCK7PT/action/replication_record"}},"created_at":"2026-06-03T01:05:28.756982+00:00","updated_at":"2026-06-03T01:05:28.756982+00:00"}