{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2020:OEVX5ZXNA7ERZDMGC4D4PPKPZQ","short_pith_number":"pith:OEVX5ZXN","schema_version":"1.0","canonical_sha256":"712b7ee6ed07c91c8d861707c7bd4fcc12e24dae59593513e547e9ca848b42f2","source":{"kind":"arxiv","id":"2012.14905","version":4},"attestation_state":"computed","paper":{"title":"Meta Learning Backpropagation And Improving It","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.NE","stat.ML"],"primary_cat":"cs.LG","authors_text":"J\\\"urgen Schmidhuber, Louis Kirsch","submitted_at":"2020-12-29T18:56:10Z","abstract_excerpt":"Many concepts have been proposed for meta learning with neural networks (NNs), e.g., NNs that learn to reprogram fast weights, Hebbian plasticity, learned learning rules, and meta recurrent NNs. Our Variable Shared Meta Learning (VSML) unifies the above and demonstrates that simple weight-sharing and sparsity in an NN is sufficient to express powerful learning algorithms (LAs) in a reusable fashion. A simple implementation of VSML where the weights of a neural network are replaced by tiny LSTMs allows for implementing the backpropagation LA solely by running in forward-mode. It can even meta l"},"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":"2012.14905","kind":"arxiv","version":4},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2020-12-29T18:56:10Z","cross_cats_sorted":["cs.AI","cs.NE","stat.ML"],"title_canon_sha256":"cb044daa8fcfd41fe69d6fe3b0f9a16e77f8eada673e5003a1cc3d1e5c0eba6d","abstract_canon_sha256":"1ce00750eb8e93da8101012bb72e8121e1e056540f9035bae3fe7e2a885e391e"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T04:04:37.466347Z","signature_b64":"3DIiHzVsl+7MWCIqhCZvF2UypyUTkkyMnFTEWmp77JVxniCSMIKLKtvlnskQaq1wbN/pFzPlrbebFKm9mUBJAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"712b7ee6ed07c91c8d861707c7bd4fcc12e24dae59593513e547e9ca848b42f2","last_reissued_at":"2026-07-05T04:04:37.465913Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T04:04:37.465913Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Meta Learning Backpropagation And Improving It","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.NE","stat.ML"],"primary_cat":"cs.LG","authors_text":"J\\\"urgen Schmidhuber, Louis Kirsch","submitted_at":"2020-12-29T18:56:10Z","abstract_excerpt":"Many concepts have been proposed for meta learning with neural networks (NNs), e.g., NNs that learn to reprogram fast weights, Hebbian plasticity, learned learning rules, and meta recurrent NNs. Our Variable Shared Meta Learning (VSML) unifies the above and demonstrates that simple weight-sharing and sparsity in an NN is sufficient to express powerful learning algorithms (LAs) in a reusable fashion. A simple implementation of VSML where the weights of a neural network are replaced by tiny LSTMs allows for implementing the backpropagation LA solely by running in forward-mode. It can even meta l"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2012.14905","kind":"arxiv","version":4},"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/2012.14905/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":"2012.14905","created_at":"2026-07-05T04:04:37.465973+00:00"},{"alias_kind":"arxiv_version","alias_value":"2012.14905v4","created_at":"2026-07-05T04:04:37.465973+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2012.14905","created_at":"2026-07-05T04:04:37.465973+00:00"},{"alias_kind":"pith_short_12","alias_value":"OEVX5ZXNA7ER","created_at":"2026-07-05T04:04:37.465973+00:00"},{"alias_kind":"pith_short_16","alias_value":"OEVX5ZXNA7ERZDMG","created_at":"2026-07-05T04:04:37.465973+00:00"},{"alias_kind":"pith_short_8","alias_value":"OEVX5ZXN","created_at":"2026-07-05T04:04:37.465973+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2507.17668","citing_title":"How Should We Meta-Learn Reinforcement Learning Algorithms?","ref_index":35,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/OEVX5ZXNA7ERZDMGC4D4PPKPZQ","json":"https://pith.science/pith/OEVX5ZXNA7ERZDMGC4D4PPKPZQ.json","graph_json":"https://pith.science/api/pith-number/OEVX5ZXNA7ERZDMGC4D4PPKPZQ/graph.json","events_json":"https://pith.science/api/pith-number/OEVX5ZXNA7ERZDMGC4D4PPKPZQ/events.json","paper":"https://pith.science/paper/OEVX5ZXN"},"agent_actions":{"view_html":"https://pith.science/pith/OEVX5ZXNA7ERZDMGC4D4PPKPZQ","download_json":"https://pith.science/pith/OEVX5ZXNA7ERZDMGC4D4PPKPZQ.json","view_paper":"https://pith.science/paper/OEVX5ZXN","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2012.14905&json=true","fetch_graph":"https://pith.science/api/pith-number/OEVX5ZXNA7ERZDMGC4D4PPKPZQ/graph.json","fetch_events":"https://pith.science/api/pith-number/OEVX5ZXNA7ERZDMGC4D4PPKPZQ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/OEVX5ZXNA7ERZDMGC4D4PPKPZQ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/OEVX5ZXNA7ERZDMGC4D4PPKPZQ/action/storage_attestation","attest_author":"https://pith.science/pith/OEVX5ZXNA7ERZDMGC4D4PPKPZQ/action/author_attestation","sign_citation":"https://pith.science/pith/OEVX5ZXNA7ERZDMGC4D4PPKPZQ/action/citation_signature","submit_replication":"https://pith.science/pith/OEVX5ZXNA7ERZDMGC4D4PPKPZQ/action/replication_record"}},"created_at":"2026-07-05T04:04:37.465973+00:00","updated_at":"2026-07-05T04:04:37.465973+00:00"}