{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2020:AQK7XYIKU75JEN74KYA2GDQ4HY","short_pith_number":"pith:AQK7XYIK","schema_version":"1.0","canonical_sha256":"0415fbe10aa7fa9237fc5601a30e1c3e1df6591f1ee6416e4f76222fb083decb","source":{"kind":"arxiv","id":"2006.07232","version":1},"attestation_state":"computed","paper":{"title":"A Practical Sparse Approximation for Real Time Recurrent Learning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.NE","stat.ML"],"primary_cat":"cs.LG","authors_text":"Alex Graves, Erich Elsen, Jacob Menick, Karen Simonyan, Simon Osindero, Utku Evci","submitted_at":"2020-06-12T14:38:15Z","abstract_excerpt":"Current methods for training recurrent neural networks are based on backpropagation through time, which requires storing a complete history of network states, and prohibits updating the weights `online' (after every timestep). Real Time Recurrent Learning (RTRL) eliminates the need for history storage and allows for online weight updates, but does so at the expense of computational costs that are quartic in the state size. This renders RTRL training intractable for all but the smallest networks, even ones that are made highly sparse.\n  We introduce the Sparse n-step Approximation (SnAp) to the"},"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":"2006.07232","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2020-06-12T14:38:15Z","cross_cats_sorted":["cs.NE","stat.ML"],"title_canon_sha256":"b61abfc64fae49c7e9fa592499f34b19896b010eb8bc2687ea8aedeeb871bc2b","abstract_canon_sha256":"2ed67a195912de3aff3871fdf0db2f051347a1be68643aa78fde0977591f0ebb"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T01:09:51.180397Z","signature_b64":"L+Fb8AAnhIScz5yCpMe8xxCsChjRL8DpzZ1ve10qYgY///ozGd5m3yIR1Ey0XAEhjeCZmk64lU0WGWlvVAu1AA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"0415fbe10aa7fa9237fc5601a30e1c3e1df6591f1ee6416e4f76222fb083decb","last_reissued_at":"2026-07-05T01:09:51.179911Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T01:09:51.179911Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"A Practical Sparse Approximation for Real Time Recurrent Learning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.NE","stat.ML"],"primary_cat":"cs.LG","authors_text":"Alex Graves, Erich Elsen, Jacob Menick, Karen Simonyan, Simon Osindero, Utku Evci","submitted_at":"2020-06-12T14:38:15Z","abstract_excerpt":"Current methods for training recurrent neural networks are based on backpropagation through time, which requires storing a complete history of network states, and prohibits updating the weights `online' (after every timestep). Real Time Recurrent Learning (RTRL) eliminates the need for history storage and allows for online weight updates, but does so at the expense of computational costs that are quartic in the state size. This renders RTRL training intractable for all but the smallest networks, even ones that are made highly sparse.\n  We introduce the Sparse n-step Approximation (SnAp) to the"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2006.07232","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/2006.07232/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":"2006.07232","created_at":"2026-07-05T01:09:51.179969+00:00"},{"alias_kind":"arxiv_version","alias_value":"2006.07232v1","created_at":"2026-07-05T01:09:51.179969+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2006.07232","created_at":"2026-07-05T01:09:51.179969+00:00"},{"alias_kind":"pith_short_12","alias_value":"AQK7XYIKU75J","created_at":"2026-07-05T01:09:51.179969+00:00"},{"alias_kind":"pith_short_16","alias_value":"AQK7XYIKU75JEN74","created_at":"2026-07-05T01:09:51.179969+00:00"},{"alias_kind":"pith_short_8","alias_value":"AQK7XYIK","created_at":"2026-07-05T01:09:51.179969+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":4,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2410.05882","citing_title":"Frame forecasting in cine MRI using the PCA respiratory motion model: comparing recurrent neural networks trained online and transformers","ref_index":26,"is_internal_anchor":false},{"citing_arxiv_id":"2605.16318","citing_title":"Investigating Action Encodings in Recurrent Neural Networks in Reinforcement Learning","ref_index":9,"is_internal_anchor":false},{"citing_arxiv_id":"2605.12763","citing_title":"State-Space NTK Collapse Near Bifurcations","ref_index":104,"is_internal_anchor":false},{"citing_arxiv_id":"2605.08746","citing_title":"The Global Empirical NTK: Self-Referential Bias and Dimensionality of Gradient Descent Learning","ref_index":104,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/AQK7XYIKU75JEN74KYA2GDQ4HY","json":"https://pith.science/pith/AQK7XYIKU75JEN74KYA2GDQ4HY.json","graph_json":"https://pith.science/api/pith-number/AQK7XYIKU75JEN74KYA2GDQ4HY/graph.json","events_json":"https://pith.science/api/pith-number/AQK7XYIKU75JEN74KYA2GDQ4HY/events.json","paper":"https://pith.science/paper/AQK7XYIK"},"agent_actions":{"view_html":"https://pith.science/pith/AQK7XYIKU75JEN74KYA2GDQ4HY","download_json":"https://pith.science/pith/AQK7XYIKU75JEN74KYA2GDQ4HY.json","view_paper":"https://pith.science/paper/AQK7XYIK","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2006.07232&json=true","fetch_graph":"https://pith.science/api/pith-number/AQK7XYIKU75JEN74KYA2GDQ4HY/graph.json","fetch_events":"https://pith.science/api/pith-number/AQK7XYIKU75JEN74KYA2GDQ4HY/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/AQK7XYIKU75JEN74KYA2GDQ4HY/action/timestamp_anchor","attest_storage":"https://pith.science/pith/AQK7XYIKU75JEN74KYA2GDQ4HY/action/storage_attestation","attest_author":"https://pith.science/pith/AQK7XYIKU75JEN74KYA2GDQ4HY/action/author_attestation","sign_citation":"https://pith.science/pith/AQK7XYIKU75JEN74KYA2GDQ4HY/action/citation_signature","submit_replication":"https://pith.science/pith/AQK7XYIKU75JEN74KYA2GDQ4HY/action/replication_record"}},"created_at":"2026-07-05T01:09:51.179969+00:00","updated_at":"2026-07-05T01:09:51.179969+00:00"}