{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:HS74GOBUCI3ALG5KK2HTIBPY3E","short_pith_number":"pith:HS74GOBU","schema_version":"1.0","canonical_sha256":"3cbfc338341236059baa568f3405f8d928a8619ae228663bf60067edca0e3172","source":{"kind":"arxiv","id":"2312.08550","version":3},"attestation_state":"computed","paper":{"title":"Harmonics of Learning: Universal Fourier Features Emerge in Invariant Networks","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","eess.SP"],"primary_cat":"cs.LG","authors_text":"Christopher Hillar, Danica Kragic, Giovanni Luca Marchetti, Sophia Sanborn","submitted_at":"2023-12-13T22:42:55Z","abstract_excerpt":"In this work, we formally prove that, under certain conditions, if a neural network is invariant to a finite group then its weights recover the Fourier transform on that group. This provides a mathematical explanation for the emergence of Fourier features -- a ubiquitous phenomenon in both biological and artificial learning systems. The results hold even for non-commutative groups, in which case the Fourier transform encodes all the irreducible unitary group representations. Our findings have consequences for the problem of symmetry discovery. Specifically, we demonstrate that the algebraic st"},"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":"2312.08550","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2023-12-13T22:42:55Z","cross_cats_sorted":["cs.AI","eess.SP"],"title_canon_sha256":"bd6318bd30a511d567115d5291c5b8e8033a8811d5bab1008f36245377500fbf","abstract_canon_sha256":"9739a9e75702582461a7fe2a04de720a31d03d3e4bd876b30bf05fdcd30e09f6"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:31:45.436848Z","signature_b64":"Tj5oJULFNRK4aYLko9LR/qZdDmZbG5Ev68B/LuTUEgBMN050kiZ/sHCp32oGS/EybH0gjoXAbk0QLG4uNnJJCg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"3cbfc338341236059baa568f3405f8d928a8619ae228663bf60067edca0e3172","last_reissued_at":"2026-07-05T08:31:45.436341Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:31:45.436341Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Harmonics of Learning: Universal Fourier Features Emerge in Invariant Networks","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","eess.SP"],"primary_cat":"cs.LG","authors_text":"Christopher Hillar, Danica Kragic, Giovanni Luca Marchetti, Sophia Sanborn","submitted_at":"2023-12-13T22:42:55Z","abstract_excerpt":"In this work, we formally prove that, under certain conditions, if a neural network is invariant to a finite group then its weights recover the Fourier transform on that group. This provides a mathematical explanation for the emergence of Fourier features -- a ubiquitous phenomenon in both biological and artificial learning systems. The results hold even for non-commutative groups, in which case the Fourier transform encodes all the irreducible unitary group representations. Our findings have consequences for the problem of symmetry discovery. Specifically, we demonstrate that the algebraic st"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2312.08550","kind":"arxiv","version":3},"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/2312.08550/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":"2312.08550","created_at":"2026-07-05T08:31:45.436396+00:00"},{"alias_kind":"arxiv_version","alias_value":"2312.08550v3","created_at":"2026-07-05T08:31:45.436396+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2312.08550","created_at":"2026-07-05T08:31:45.436396+00:00"},{"alias_kind":"pith_short_12","alias_value":"HS74GOBUCI3A","created_at":"2026-07-05T08:31:45.436396+00:00"},{"alias_kind":"pith_short_16","alias_value":"HS74GOBUCI3ALG5K","created_at":"2026-07-05T08:31:45.436396+00:00"},{"alias_kind":"pith_short_8","alias_value":"HS74GOBU","created_at":"2026-07-05T08:31:45.436396+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2508.10490","citing_title":"On the Complexity-Faithfulness Trade-off of Gradient-Based Explanations","ref_index":40,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/HS74GOBUCI3ALG5KK2HTIBPY3E","json":"https://pith.science/pith/HS74GOBUCI3ALG5KK2HTIBPY3E.json","graph_json":"https://pith.science/api/pith-number/HS74GOBUCI3ALG5KK2HTIBPY3E/graph.json","events_json":"https://pith.science/api/pith-number/HS74GOBUCI3ALG5KK2HTIBPY3E/events.json","paper":"https://pith.science/paper/HS74GOBU"},"agent_actions":{"view_html":"https://pith.science/pith/HS74GOBUCI3ALG5KK2HTIBPY3E","download_json":"https://pith.science/pith/HS74GOBUCI3ALG5KK2HTIBPY3E.json","view_paper":"https://pith.science/paper/HS74GOBU","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2312.08550&json=true","fetch_graph":"https://pith.science/api/pith-number/HS74GOBUCI3ALG5KK2HTIBPY3E/graph.json","fetch_events":"https://pith.science/api/pith-number/HS74GOBUCI3ALG5KK2HTIBPY3E/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/HS74GOBUCI3ALG5KK2HTIBPY3E/action/timestamp_anchor","attest_storage":"https://pith.science/pith/HS74GOBUCI3ALG5KK2HTIBPY3E/action/storage_attestation","attest_author":"https://pith.science/pith/HS74GOBUCI3ALG5KK2HTIBPY3E/action/author_attestation","sign_citation":"https://pith.science/pith/HS74GOBUCI3ALG5KK2HTIBPY3E/action/citation_signature","submit_replication":"https://pith.science/pith/HS74GOBUCI3ALG5KK2HTIBPY3E/action/replication_record"}},"created_at":"2026-07-05T08:31:45.436396+00:00","updated_at":"2026-07-05T08:31:45.436396+00:00"}