{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2018:NNBSYPG6H7IAOI7RCAP3ELAIKW","short_pith_number":"pith:NNBSYPG6","schema_version":"1.0","canonical_sha256":"6b432c3cde3fd00723f1101fb22c0855bdf5a8d898eda906afe88f8ba45c1549","source":{"kind":"arxiv","id":"1806.08734","version":3},"attestation_state":"computed","paper":{"title":"On the Spectral Bias of Neural Networks","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"stat.ML","authors_text":"Aaron Courville, Aristide Baratin, Devansh Arpit, Felix Draxler, Fred A. Hamprecht, Min Lin, Nasim Rahaman, Yoshua Bengio","submitted_at":"2018-06-22T15:39:05Z","abstract_excerpt":"Neural networks are known to be a class of highly expressive functions able to fit even random input-output mappings with $100\\%$ accuracy. In this work, we present properties of neural networks that complement this aspect of expressivity. By using tools from Fourier analysis, we show that deep ReLU networks are biased towards low frequency functions, meaning that they cannot have local fluctuations without affecting their global behavior. Intuitively, this property is in line with the observation that over-parameterized networks find simple patterns that generalize across data samples. We als"},"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":"1806.08734","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"stat.ML","submitted_at":"2018-06-22T15:39:05Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"01010d8f1349482cd1ec3aa97406bd727913fd832427480b8f45fc617cfdca76","abstract_canon_sha256":"83886ec7e318a93412e6b84890d403e2c3c332a537d071b674dec7ffbba1e88d"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-17T23:44:38.177286Z","signature_b64":"1XK3Ec/qLcYKGv7OgE8qNk2OM/jAzdM71vf0d+tGjie7CmWckdtIwt1zJ94xW6VUbEwftWSFrGpEEbeM2JpgCA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"6b432c3cde3fd00723f1101fb22c0855bdf5a8d898eda906afe88f8ba45c1549","last_reissued_at":"2026-05-17T23:44:38.176928Z","signature_status":"signed_v1","first_computed_at":"2026-05-17T23:44:38.176928Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"On the Spectral Bias of Neural Networks","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"stat.ML","authors_text":"Aaron Courville, Aristide Baratin, Devansh Arpit, Felix Draxler, Fred A. Hamprecht, Min Lin, Nasim Rahaman, Yoshua Bengio","submitted_at":"2018-06-22T15:39:05Z","abstract_excerpt":"Neural networks are known to be a class of highly expressive functions able to fit even random input-output mappings with $100\\%$ accuracy. In this work, we present properties of neural networks that complement this aspect of expressivity. By using tools from Fourier analysis, we show that deep ReLU networks are biased towards low frequency functions, meaning that they cannot have local fluctuations without affecting their global behavior. Intuitively, this property is in line with the observation that over-parameterized networks find simple patterns that generalize across data samples. We als"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1806.08734","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":""},"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":"1806.08734","created_at":"2026-05-17T23:44:38.176989+00:00"},{"alias_kind":"arxiv_version","alias_value":"1806.08734v3","created_at":"2026-05-17T23:44:38.176989+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1806.08734","created_at":"2026-05-17T23:44:38.176989+00:00"},{"alias_kind":"pith_short_12","alias_value":"NNBSYPG6H7IA","created_at":"2026-05-18T12:32:40.477152+00:00"},{"alias_kind":"pith_short_16","alias_value":"NNBSYPG6H7IAOI7R","created_at":"2026-05-18T12:32:40.477152+00:00"},{"alias_kind":"pith_short_8","alias_value":"NNBSYPG6","created_at":"2026-05-18T12:32:40.477152+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":12,"internal_anchor_count":5,"sample":[{"citing_arxiv_id":"1906.09235","citing_title":"Theory of the Frequency Principle for General Deep Neural Networks","ref_index":8,"is_internal_anchor":true},{"citing_arxiv_id":"2407.12173","citing_title":"Beta Sampling is All You Need: Efficient Image Generation Strategy for Diffusion Models using Stepwise Spectral Analysis","ref_index":22,"is_internal_anchor":true},{"citing_arxiv_id":"2509.12344","citing_title":"FEDONet : Fourier-Embedded DeepONet for Spectrally Accurate Operator Learning","ref_index":51,"is_internal_anchor":true},{"citing_arxiv_id":"2603.20120","citing_title":"Deep learning-based phase-field modelling of brittle fracture in anisotropic media","ref_index":46,"is_internal_anchor":true},{"citing_arxiv_id":"2605.08512","citing_title":"MoMo: Conditioned Contrastive Representation Learning for Preference-Modulated Planning","ref_index":64,"is_internal_anchor":true},{"citing_arxiv_id":"2605.08512","citing_title":"MoMo: Conditioned Contrastive Representation Learning for Preference-Modulated Planning","ref_index":64,"is_internal_anchor":false},{"citing_arxiv_id":"2605.01288","citing_title":"A Theory of Saddle Escape in Deep Nonlinear Networks","ref_index":34,"is_internal_anchor":false},{"citing_arxiv_id":"2604.18686","citing_title":"Neural Spectral Bias and Conformal Correlators I: Introduction and Applications","ref_index":20,"is_internal_anchor":false},{"citing_arxiv_id":"2604.18673","citing_title":"Neural Networks Reveal a Universal Bias in Conformal Correlators","ref_index":24,"is_internal_anchor":false},{"citing_arxiv_id":"2605.01288","citing_title":"A Theory of Saddle Escape in Deep Nonlinear Networks","ref_index":34,"is_internal_anchor":false},{"citing_arxiv_id":"2604.05394","citing_title":"Neural Assistive Impulses: Synthesizing Exaggerated Motions for Physics-based Characters","ref_index":22,"is_internal_anchor":false},{"citing_arxiv_id":"2605.00510","citing_title":"Scale-Aware Adversarial Analysis: A Diagnostic for Generative AI in Multiscale Complex Systems","ref_index":83,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/NNBSYPG6H7IAOI7RCAP3ELAIKW","json":"https://pith.science/pith/NNBSYPG6H7IAOI7RCAP3ELAIKW.json","graph_json":"https://pith.science/api/pith-number/NNBSYPG6H7IAOI7RCAP3ELAIKW/graph.json","events_json":"https://pith.science/api/pith-number/NNBSYPG6H7IAOI7RCAP3ELAIKW/events.json","paper":"https://pith.science/paper/NNBSYPG6"},"agent_actions":{"view_html":"https://pith.science/pith/NNBSYPG6H7IAOI7RCAP3ELAIKW","download_json":"https://pith.science/pith/NNBSYPG6H7IAOI7RCAP3ELAIKW.json","view_paper":"https://pith.science/paper/NNBSYPG6","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=1806.08734&json=true","fetch_graph":"https://pith.science/api/pith-number/NNBSYPG6H7IAOI7RCAP3ELAIKW/graph.json","fetch_events":"https://pith.science/api/pith-number/NNBSYPG6H7IAOI7RCAP3ELAIKW/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/NNBSYPG6H7IAOI7RCAP3ELAIKW/action/timestamp_anchor","attest_storage":"https://pith.science/pith/NNBSYPG6H7IAOI7RCAP3ELAIKW/action/storage_attestation","attest_author":"https://pith.science/pith/NNBSYPG6H7IAOI7RCAP3ELAIKW/action/author_attestation","sign_citation":"https://pith.science/pith/NNBSYPG6H7IAOI7RCAP3ELAIKW/action/citation_signature","submit_replication":"https://pith.science/pith/NNBSYPG6H7IAOI7RCAP3ELAIKW/action/replication_record"}},"created_at":"2026-05-17T23:44:38.176989+00:00","updated_at":"2026-05-17T23:44:38.176989+00:00"}