{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2015:6OPTND23OROUTWJQPMDTXNYE3A","short_pith_number":"pith:6OPTND23","schema_version":"1.0","canonical_sha256":"f39f368f5b745d49d9307b073bb704d8193463e7252088f19eb8f5b56b04167e","source":{"kind":"arxiv","id":"1506.08700","version":4},"attestation_state":"computed","paper":{"title":"Dropout as data augmentation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"stat.ML","authors_text":"Kishore Konda, Pascal Vincent, Roland Memisevic, Xavier Bouthillier","submitted_at":"2015-06-29T15:55:45Z","abstract_excerpt":"Dropout is typically interpreted as bagging a large number of models sharing parameters. We show that using dropout in a network can also be interpreted as a kind of data augmentation in the input space without domain knowledge. We present an approach to projecting the dropout noise within a network back into the input space, thereby generating augmented versions of the training data, and we show that training a deterministic network on the augmented samples yields similar results. Finally, we propose a new dropout noise scheme based on our observations and show that it improves dropout result"},"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":"1506.08700","kind":"arxiv","version":4},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"stat.ML","submitted_at":"2015-06-29T15:55:45Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"9ce685d1129e2c7f86d01848482592e739e6679ad676d95b3498b70d2c40dd3a","abstract_canon_sha256":"036ebfaa2067eb775e726dcf9cea636ab3256d276496cc3997e2473bfbebdf49"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-18T01:23:12.131991Z","signature_b64":"bjKkRNAf4SSsxRNLnFKA9MHm1jLcO+aFqC8DIzNbiVP6DGNUigKK356FOtmtCP8i84AFrrMsdL+x73ROz5uTCw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"f39f368f5b745d49d9307b073bb704d8193463e7252088f19eb8f5b56b04167e","last_reissued_at":"2026-05-18T01:23:12.131470Z","signature_status":"signed_v1","first_computed_at":"2026-05-18T01:23:12.131470Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Dropout as data augmentation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"stat.ML","authors_text":"Kishore Konda, Pascal Vincent, Roland Memisevic, Xavier Bouthillier","submitted_at":"2015-06-29T15:55:45Z","abstract_excerpt":"Dropout is typically interpreted as bagging a large number of models sharing parameters. We show that using dropout in a network can also be interpreted as a kind of data augmentation in the input space without domain knowledge. We present an approach to projecting the dropout noise within a network back into the input space, thereby generating augmented versions of the training data, and we show that training a deterministic network on the augmented samples yields similar results. Finally, we propose a new dropout noise scheme based on our observations and show that it improves dropout result"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1506.08700","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":""},"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":"1506.08700","created_at":"2026-05-18T01:23:12.131550+00:00"},{"alias_kind":"arxiv_version","alias_value":"1506.08700v4","created_at":"2026-05-18T01:23:12.131550+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1506.08700","created_at":"2026-05-18T01:23:12.131550+00:00"},{"alias_kind":"pith_short_12","alias_value":"6OPTND23OROU","created_at":"2026-05-18T12:29:07.941421+00:00"},{"alias_kind":"pith_short_16","alias_value":"6OPTND23OROUTWJQ","created_at":"2026-05-18T12:29:07.941421+00:00"},{"alias_kind":"pith_short_8","alias_value":"6OPTND23","created_at":"2026-05-18T12:29:07.941421+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2604.17384","citing_title":"Towards a Data-Parameter Correspondence for LLMs: A Preliminary Discussion","ref_index":22,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/6OPTND23OROUTWJQPMDTXNYE3A","json":"https://pith.science/pith/6OPTND23OROUTWJQPMDTXNYE3A.json","graph_json":"https://pith.science/api/pith-number/6OPTND23OROUTWJQPMDTXNYE3A/graph.json","events_json":"https://pith.science/api/pith-number/6OPTND23OROUTWJQPMDTXNYE3A/events.json","paper":"https://pith.science/paper/6OPTND23"},"agent_actions":{"view_html":"https://pith.science/pith/6OPTND23OROUTWJQPMDTXNYE3A","download_json":"https://pith.science/pith/6OPTND23OROUTWJQPMDTXNYE3A.json","view_paper":"https://pith.science/paper/6OPTND23","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=1506.08700&json=true","fetch_graph":"https://pith.science/api/pith-number/6OPTND23OROUTWJQPMDTXNYE3A/graph.json","fetch_events":"https://pith.science/api/pith-number/6OPTND23OROUTWJQPMDTXNYE3A/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/6OPTND23OROUTWJQPMDTXNYE3A/action/timestamp_anchor","attest_storage":"https://pith.science/pith/6OPTND23OROUTWJQPMDTXNYE3A/action/storage_attestation","attest_author":"https://pith.science/pith/6OPTND23OROUTWJQPMDTXNYE3A/action/author_attestation","sign_citation":"https://pith.science/pith/6OPTND23OROUTWJQPMDTXNYE3A/action/citation_signature","submit_replication":"https://pith.science/pith/6OPTND23OROUTWJQPMDTXNYE3A/action/replication_record"}},"created_at":"2026-05-18T01:23:12.131550+00:00","updated_at":"2026-05-18T01:23:12.131550+00:00"}