{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:QAEVMT75Q3P6FE4E4D4NIJLBWT","short_pith_number":"pith:QAEVMT75","schema_version":"1.0","canonical_sha256":"8009564ffd86dfe29384e0f8d42561b4f398ba6307dabfcb7b0cd470a49d57da","source":{"kind":"arxiv","id":"2410.23467","version":2},"attestation_state":"computed","paper":{"title":"Gradient-free training of recurrent neural networks","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.NA","math.NA"],"primary_cat":"cs.LG","authors_text":"Ana Cukarska, Erik Lien Bolager, Felix Dietrich, Iryna Burak, Zahra Monfared","submitted_at":"2024-10-30T21:24:34Z","abstract_excerpt":"Recurrent neural networks are a successful neural architecture for many time-dependent problems, including time series analysis, forecasting, and modeling of dynamical systems. Training such networks with backpropagation through time is a notoriously difficult problem because their loss gradients tend to explode or vanish. In this contribution, we introduce a computational approach to construct all weights and biases of a recurrent neural network without using gradient-based methods. The approach is based on a combination of random feature networks and Koopman operator theory for dynamical sys"},"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":"2410.23467","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-10-30T21:24:34Z","cross_cats_sorted":["cs.NA","math.NA"],"title_canon_sha256":"b96acc8f530798343b3925abe5aaf04e8bd81646a19226c5a44b4b1f07609e09","abstract_canon_sha256":"35be85174bfb2f12fcc260f0d52782d6de6c98b70f5ac50de14267ce9eea935a"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:06:45.418586Z","signature_b64":"E8rPsqpbAZ71wKW3aJ3VjhBUHSeVpE8+uWnFti2lPnJ415Ig/Z5ALgGMgbNf96tHD2EqI/yr5FStgOaDgd9YAQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"8009564ffd86dfe29384e0f8d42561b4f398ba6307dabfcb7b0cd470a49d57da","last_reissued_at":"2026-07-05T10:06:45.418055Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:06:45.418055Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Gradient-free training of recurrent neural networks","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.NA","math.NA"],"primary_cat":"cs.LG","authors_text":"Ana Cukarska, Erik Lien Bolager, Felix Dietrich, Iryna Burak, Zahra Monfared","submitted_at":"2024-10-30T21:24:34Z","abstract_excerpt":"Recurrent neural networks are a successful neural architecture for many time-dependent problems, including time series analysis, forecasting, and modeling of dynamical systems. Training such networks with backpropagation through time is a notoriously difficult problem because their loss gradients tend to explode or vanish. In this contribution, we introduce a computational approach to construct all weights and biases of a recurrent neural network without using gradient-based methods. The approach is based on a combination of random feature networks and Koopman operator theory for dynamical sys"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2410.23467","kind":"arxiv","version":2},"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/2410.23467/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":"2410.23467","created_at":"2026-07-05T10:06:45.418129+00:00"},{"alias_kind":"arxiv_version","alias_value":"2410.23467v2","created_at":"2026-07-05T10:06:45.418129+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2410.23467","created_at":"2026-07-05T10:06:45.418129+00:00"},{"alias_kind":"pith_short_12","alias_value":"QAEVMT75Q3P6","created_at":"2026-07-05T10:06:45.418129+00:00"},{"alias_kind":"pith_short_16","alias_value":"QAEVMT75Q3P6FE4E","created_at":"2026-07-05T10:06:45.418129+00:00"},{"alias_kind":"pith_short_8","alias_value":"QAEVMT75","created_at":"2026-07-05T10:06:45.418129+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":2,"sample":[{"citing_arxiv_id":"2506.06558","citing_title":"Rapid training of Hamiltonian graph networks using random features","ref_index":5,"is_internal_anchor":true},{"citing_arxiv_id":"2604.18402","citing_title":"Adaptive Kernel Selection for Kernelized Diffusion Maps","ref_index":290,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/QAEVMT75Q3P6FE4E4D4NIJLBWT","json":"https://pith.science/pith/QAEVMT75Q3P6FE4E4D4NIJLBWT.json","graph_json":"https://pith.science/api/pith-number/QAEVMT75Q3P6FE4E4D4NIJLBWT/graph.json","events_json":"https://pith.science/api/pith-number/QAEVMT75Q3P6FE4E4D4NIJLBWT/events.json","paper":"https://pith.science/paper/QAEVMT75"},"agent_actions":{"view_html":"https://pith.science/pith/QAEVMT75Q3P6FE4E4D4NIJLBWT","download_json":"https://pith.science/pith/QAEVMT75Q3P6FE4E4D4NIJLBWT.json","view_paper":"https://pith.science/paper/QAEVMT75","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2410.23467&json=true","fetch_graph":"https://pith.science/api/pith-number/QAEVMT75Q3P6FE4E4D4NIJLBWT/graph.json","fetch_events":"https://pith.science/api/pith-number/QAEVMT75Q3P6FE4E4D4NIJLBWT/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/QAEVMT75Q3P6FE4E4D4NIJLBWT/action/timestamp_anchor","attest_storage":"https://pith.science/pith/QAEVMT75Q3P6FE4E4D4NIJLBWT/action/storage_attestation","attest_author":"https://pith.science/pith/QAEVMT75Q3P6FE4E4D4NIJLBWT/action/author_attestation","sign_citation":"https://pith.science/pith/QAEVMT75Q3P6FE4E4D4NIJLBWT/action/citation_signature","submit_replication":"https://pith.science/pith/QAEVMT75Q3P6FE4E4D4NIJLBWT/action/replication_record"}},"created_at":"2026-07-05T10:06:45.418129+00:00","updated_at":"2026-07-05T10:06:45.418129+00:00"}