{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:5VE5RS2K6YWT3Y6B2FKIWUDUAB","short_pith_number":"pith:5VE5RS2K","schema_version":"1.0","canonical_sha256":"ed49d8cb4af62d3de3c1d1548b507400419aa22562301b0c08f4d169e906888b","source":{"kind":"arxiv","id":"2402.06184","version":1},"attestation_state":"computed","paper":{"title":"The boundary of neural network trainability is fractal","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.NE","nlin.CD"],"primary_cat":"cs.LG","authors_text":"Jascha Sohl-Dickstein","submitted_at":"2024-02-09T04:46:48Z","abstract_excerpt":"Some fractals -- for instance those associated with the Mandelbrot and quadratic Julia sets -- are computed by iterating a function, and identifying the boundary between hyperparameters for which the resulting series diverges or remains bounded. Neural network training similarly involves iterating an update function (e.g. repeated steps of gradient descent), can result in convergent or divergent behavior, and can be extremely sensitive to small changes in hyperparameters. Motivated by these similarities, we experimentally examine the boundary between neural network hyperparameters that lead to"},"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":"2402.06184","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-02-09T04:46:48Z","cross_cats_sorted":["cs.NE","nlin.CD"],"title_canon_sha256":"2d4deec02e9096e9c56f43bef1e3e70120bccd50e6cb1cf3b7a7e8c47e91f878","abstract_canon_sha256":"c215c8261280fd9ab3926da028177c49bec7b7cf5c90e87837f08d6ac244d3e3"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T07:43:19.906076Z","signature_b64":"+JnbEPWZqwrk8oyA6/glhlD8/9WT24rctirVfFIFhXVAJlV6qNtrDa5zONeUfBaCpbEnZrhcUHE7LNJHZNwMAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"ed49d8cb4af62d3de3c1d1548b507400419aa22562301b0c08f4d169e906888b","last_reissued_at":"2026-07-05T07:43:19.905626Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T07:43:19.905626Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"The boundary of neural network trainability is fractal","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.NE","nlin.CD"],"primary_cat":"cs.LG","authors_text":"Jascha Sohl-Dickstein","submitted_at":"2024-02-09T04:46:48Z","abstract_excerpt":"Some fractals -- for instance those associated with the Mandelbrot and quadratic Julia sets -- are computed by iterating a function, and identifying the boundary between hyperparameters for which the resulting series diverges or remains bounded. Neural network training similarly involves iterating an update function (e.g. repeated steps of gradient descent), can result in convergent or divergent behavior, and can be extremely sensitive to small changes in hyperparameters. Motivated by these similarities, we experimentally examine the boundary between neural network hyperparameters that lead to"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2402.06184","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/2402.06184/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":"2402.06184","created_at":"2026-07-05T07:43:19.905680+00:00"},{"alias_kind":"arxiv_version","alias_value":"2402.06184v1","created_at":"2026-07-05T07:43:19.905680+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2402.06184","created_at":"2026-07-05T07:43:19.905680+00:00"},{"alias_kind":"pith_short_12","alias_value":"5VE5RS2K6YWT","created_at":"2026-07-05T07:43:19.905680+00:00"},{"alias_kind":"pith_short_16","alias_value":"5VE5RS2K6YWT3Y6B","created_at":"2026-07-05T07:43:19.905680+00:00"},{"alias_kind":"pith_short_8","alias_value":"5VE5RS2K","created_at":"2026-07-05T07:43:19.905680+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2506.13234","citing_title":"The Butterfly Effect: Neural Network Training Trajectories Are Highly Sensitive to Initial Conditions","ref_index":61,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/5VE5RS2K6YWT3Y6B2FKIWUDUAB","json":"https://pith.science/pith/5VE5RS2K6YWT3Y6B2FKIWUDUAB.json","graph_json":"https://pith.science/api/pith-number/5VE5RS2K6YWT3Y6B2FKIWUDUAB/graph.json","events_json":"https://pith.science/api/pith-number/5VE5RS2K6YWT3Y6B2FKIWUDUAB/events.json","paper":"https://pith.science/paper/5VE5RS2K"},"agent_actions":{"view_html":"https://pith.science/pith/5VE5RS2K6YWT3Y6B2FKIWUDUAB","download_json":"https://pith.science/pith/5VE5RS2K6YWT3Y6B2FKIWUDUAB.json","view_paper":"https://pith.science/paper/5VE5RS2K","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2402.06184&json=true","fetch_graph":"https://pith.science/api/pith-number/5VE5RS2K6YWT3Y6B2FKIWUDUAB/graph.json","fetch_events":"https://pith.science/api/pith-number/5VE5RS2K6YWT3Y6B2FKIWUDUAB/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/5VE5RS2K6YWT3Y6B2FKIWUDUAB/action/timestamp_anchor","attest_storage":"https://pith.science/pith/5VE5RS2K6YWT3Y6B2FKIWUDUAB/action/storage_attestation","attest_author":"https://pith.science/pith/5VE5RS2K6YWT3Y6B2FKIWUDUAB/action/author_attestation","sign_citation":"https://pith.science/pith/5VE5RS2K6YWT3Y6B2FKIWUDUAB/action/citation_signature","submit_replication":"https://pith.science/pith/5VE5RS2K6YWT3Y6B2FKIWUDUAB/action/replication_record"}},"created_at":"2026-07-05T07:43:19.905680+00:00","updated_at":"2026-07-05T07:43:19.905680+00:00"}