{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2016:NGCW533YUUPOXVLS6BAOPPV3ZJ","short_pith_number":"pith:NGCW533Y","schema_version":"1.0","canonical_sha256":"69856eef78a51eebd572f040e7bebbca7565356a293cb498234fd6dc9c5caff1","source":{"kind":"arxiv","id":"1608.05343","version":2},"attestation_state":"computed","paper":{"title":"Decoupled Neural Interfaces using Synthetic Gradients","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Alex Graves, David Silver, Koray Kavukcuoglu, Max Jaderberg, Oriol Vinyals, Simon Osindero, Wojciech Marian Czarnecki","submitted_at":"2016-08-18T17:29:09Z","abstract_excerpt":"Training directed neural networks typically requires forward-propagating data through a computation graph, followed by backpropagating error signal, to produce weight updates. All layers, or more generally, modules, of the network are therefore locked, in the sense that they must wait for the remainder of the network to execute forwards and propagate error backwards before they can be updated. In this work we break this constraint by decoupling modules by introducing a model of the future computation of the network graph. These models predict what the result of the modelled subgraph will produ"},"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":"1608.05343","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2016-08-18T17:29:09Z","cross_cats_sorted":[],"title_canon_sha256":"919c5ee0b86616f17b71f5c847ed81188dc5b76cb1df6ca2aa076689e7d3c1f3","abstract_canon_sha256":"019acc131b2e292a1b9326f150c9ab17551e1e29503ac8b5ae419dd0d8cf2643"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-18T00:41:07.968849Z","signature_b64":"ue6VURpJNbTuh2uFTrT0uHP0Y8lsnkctvKUNIccKJCouBQKMvnOHsfE5KsxsEjFq98TO0dl9W8gZCgY8x4/wDQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"69856eef78a51eebd572f040e7bebbca7565356a293cb498234fd6dc9c5caff1","last_reissued_at":"2026-05-18T00:41:07.968279Z","signature_status":"signed_v1","first_computed_at":"2026-05-18T00:41:07.968279Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Decoupled Neural Interfaces using Synthetic Gradients","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Alex Graves, David Silver, Koray Kavukcuoglu, Max Jaderberg, Oriol Vinyals, Simon Osindero, Wojciech Marian Czarnecki","submitted_at":"2016-08-18T17:29:09Z","abstract_excerpt":"Training directed neural networks typically requires forward-propagating data through a computation graph, followed by backpropagating error signal, to produce weight updates. All layers, or more generally, modules, of the network are therefore locked, in the sense that they must wait for the remainder of the network to execute forwards and propagate error backwards before they can be updated. In this work we break this constraint by decoupling modules by introducing a model of the future computation of the network graph. These models predict what the result of the modelled subgraph will produ"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1608.05343","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":""},"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":"1608.05343","created_at":"2026-05-18T00:41:07.968391+00:00"},{"alias_kind":"arxiv_version","alias_value":"1608.05343v2","created_at":"2026-05-18T00:41:07.968391+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1608.05343","created_at":"2026-05-18T00:41:07.968391+00:00"},{"alias_kind":"pith_short_12","alias_value":"NGCW533YUUPO","created_at":"2026-05-18T12:30:32.724797+00:00"},{"alias_kind":"pith_short_16","alias_value":"NGCW533YUUPOXVLS","created_at":"2026-05-18T12:30:32.724797+00:00"},{"alias_kind":"pith_short_8","alias_value":"NGCW533Y","created_at":"2026-05-18T12:30:32.724797+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2206.07682","citing_title":"Emergent Abilities of Large Language Models","ref_index":40,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/NGCW533YUUPOXVLS6BAOPPV3ZJ","json":"https://pith.science/pith/NGCW533YUUPOXVLS6BAOPPV3ZJ.json","graph_json":"https://pith.science/api/pith-number/NGCW533YUUPOXVLS6BAOPPV3ZJ/graph.json","events_json":"https://pith.science/api/pith-number/NGCW533YUUPOXVLS6BAOPPV3ZJ/events.json","paper":"https://pith.science/paper/NGCW533Y"},"agent_actions":{"view_html":"https://pith.science/pith/NGCW533YUUPOXVLS6BAOPPV3ZJ","download_json":"https://pith.science/pith/NGCW533YUUPOXVLS6BAOPPV3ZJ.json","view_paper":"https://pith.science/paper/NGCW533Y","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=1608.05343&json=true","fetch_graph":"https://pith.science/api/pith-number/NGCW533YUUPOXVLS6BAOPPV3ZJ/graph.json","fetch_events":"https://pith.science/api/pith-number/NGCW533YUUPOXVLS6BAOPPV3ZJ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/NGCW533YUUPOXVLS6BAOPPV3ZJ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/NGCW533YUUPOXVLS6BAOPPV3ZJ/action/storage_attestation","attest_author":"https://pith.science/pith/NGCW533YUUPOXVLS6BAOPPV3ZJ/action/author_attestation","sign_citation":"https://pith.science/pith/NGCW533YUUPOXVLS6BAOPPV3ZJ/action/citation_signature","submit_replication":"https://pith.science/pith/NGCW533YUUPOXVLS6BAOPPV3ZJ/action/replication_record"}},"created_at":"2026-05-18T00:41:07.968391+00:00","updated_at":"2026-05-18T00:41:07.968391+00:00"}