{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2013:6IT34CNWDA4I2F7LICHSHMMMQ6","short_pith_number":"pith:6IT34CNW","schema_version":"1.0","canonical_sha256":"f227be09b618388d17eb408f23b18c87a8da49b3582a4a6ce0108a85e8304ba2","source":{"kind":"arxiv","id":"1301.3476","version":3},"attestation_state":"computed","paper":{"title":"Pushing Stochastic Gradient towards Second-Order Methods -- Backpropagation Learning with Transformations in Nonlinearities","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CV","stat.ML"],"primary_cat":"cs.LG","authors_text":"Harri Valpola, Tapani Raiko, Tommi Vatanen, Yann LeCun","submitted_at":"2013-01-15T20:21:54Z","abstract_excerpt":"Recently, we proposed to transform the outputs of each hidden neuron in a multi-layer perceptron network to have zero output and zero slope on average, and use separate shortcut connections to model the linear dependencies instead. We continue the work by firstly introducing a third transformation to normalize the scale of the outputs of each hidden neuron, and secondly by analyzing the connections to second order optimization methods. We show that the transformations make a simple stochastic gradient behave closer to second-order optimization methods and thus speed up learning. This is shown "},"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":"1301.3476","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2013-01-15T20:21:54Z","cross_cats_sorted":["cs.CV","stat.ML"],"title_canon_sha256":"a892b6e71b5a613a81ca0313731829361ef241acb2d3e22d018678ae0fb70e86","abstract_canon_sha256":"12afe81e9c734f0ec4c544bcc0b35a2396010bca63cb1c0c4dd0c392cdb587e1"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-18T03:31:20.417877Z","signature_b64":"vOgkhhZ+/JdOfBizBNDmRZcogCUhCnaWJNBIGQelfXDJfdbPCFaHnvDfGcq6OGz5RC755sbck2XTcArL+3QUAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"f227be09b618388d17eb408f23b18c87a8da49b3582a4a6ce0108a85e8304ba2","last_reissued_at":"2026-05-18T03:31:20.417086Z","signature_status":"signed_v1","first_computed_at":"2026-05-18T03:31:20.417086Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Pushing Stochastic Gradient towards Second-Order Methods -- Backpropagation Learning with Transformations in Nonlinearities","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CV","stat.ML"],"primary_cat":"cs.LG","authors_text":"Harri Valpola, Tapani Raiko, Tommi Vatanen, Yann LeCun","submitted_at":"2013-01-15T20:21:54Z","abstract_excerpt":"Recently, we proposed to transform the outputs of each hidden neuron in a multi-layer perceptron network to have zero output and zero slope on average, and use separate shortcut connections to model the linear dependencies instead. We continue the work by firstly introducing a third transformation to normalize the scale of the outputs of each hidden neuron, and secondly by analyzing the connections to second order optimization methods. We show that the transformations make a simple stochastic gradient behave closer to second-order optimization methods and thus speed up learning. This is shown "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1301.3476","kind":"arxiv","version":3},"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":"1301.3476","created_at":"2026-05-18T03:31:20.417236+00:00"},{"alias_kind":"arxiv_version","alias_value":"1301.3476v3","created_at":"2026-05-18T03:31:20.417236+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1301.3476","created_at":"2026-05-18T03:31:20.417236+00:00"},{"alias_kind":"pith_short_12","alias_value":"6IT34CNWDA4I","created_at":"2026-05-18T12:27:36.564083+00:00"},{"alias_kind":"pith_short_16","alias_value":"6IT34CNWDA4I2F7L","created_at":"2026-05-18T12:27:36.564083+00:00"},{"alias_kind":"pith_short_8","alias_value":"6IT34CNW","created_at":"2026-05-18T12:27:36.564083+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/6IT34CNWDA4I2F7LICHSHMMMQ6","json":"https://pith.science/pith/6IT34CNWDA4I2F7LICHSHMMMQ6.json","graph_json":"https://pith.science/api/pith-number/6IT34CNWDA4I2F7LICHSHMMMQ6/graph.json","events_json":"https://pith.science/api/pith-number/6IT34CNWDA4I2F7LICHSHMMMQ6/events.json","paper":"https://pith.science/paper/6IT34CNW"},"agent_actions":{"view_html":"https://pith.science/pith/6IT34CNWDA4I2F7LICHSHMMMQ6","download_json":"https://pith.science/pith/6IT34CNWDA4I2F7LICHSHMMMQ6.json","view_paper":"https://pith.science/paper/6IT34CNW","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=1301.3476&json=true","fetch_graph":"https://pith.science/api/pith-number/6IT34CNWDA4I2F7LICHSHMMMQ6/graph.json","fetch_events":"https://pith.science/api/pith-number/6IT34CNWDA4I2F7LICHSHMMMQ6/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/6IT34CNWDA4I2F7LICHSHMMMQ6/action/timestamp_anchor","attest_storage":"https://pith.science/pith/6IT34CNWDA4I2F7LICHSHMMMQ6/action/storage_attestation","attest_author":"https://pith.science/pith/6IT34CNWDA4I2F7LICHSHMMMQ6/action/author_attestation","sign_citation":"https://pith.science/pith/6IT34CNWDA4I2F7LICHSHMMMQ6/action/citation_signature","submit_replication":"https://pith.science/pith/6IT34CNWDA4I2F7LICHSHMMMQ6/action/replication_record"}},"created_at":"2026-05-18T03:31:20.417236+00:00","updated_at":"2026-05-18T03:31:20.417236+00:00"}