{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2017:5LR5GJA3WHNFQTWWNTBY2A5JYP","short_pith_number":"pith:5LR5GJA3","schema_version":"1.0","canonical_sha256":"eae3d3241bb1da584ed66cc38d03a9c3f92f96fe1414ce0f002af914f1a48a45","source":{"kind":"arxiv","id":"1711.00123","version":3},"attestation_state":"computed","paper":{"title":"Backpropagation through the Void: Optimizing control variates for black-box gradient estimation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Dami Choi, David Duvenaud, Geoffrey Roeder, Will Grathwohl, Yuhuai Wu","submitted_at":"2017-10-31T21:58:13Z","abstract_excerpt":"Gradient-based optimization is the foundation of deep learning and reinforcement learning. Even when the mechanism being optimized is unknown or not differentiable, optimization using high-variance or biased gradient estimates is still often the best strategy. We introduce a general framework for learning low-variance, unbiased gradient estimators for black-box functions of random variables. Our method uses gradients of a neural network trained jointly with model parameters or policies, and is applicable in both discrete and continuous settings. We demonstrate this framework for training discr"},"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":"1711.00123","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2017-10-31T21:58:13Z","cross_cats_sorted":[],"title_canon_sha256":"f51f987287647db3cab345093a29e297bdbf7f5deebd0a8b7158732ac1aa32fc","abstract_canon_sha256":"853abf20b149955c8efe04b082832a0292cadd30781781a965453672fb7484bc"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-18T00:22:39.291714Z","signature_b64":"TNO2KP6gxzwBcLWJt3pTQT0b9P5p1iRVU+KZ1mIxcpHZ7ZmoxwYWfIxgjSDSjrbOzkI/DNbNQ8tSzhs0cwGjBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"eae3d3241bb1da584ed66cc38d03a9c3f92f96fe1414ce0f002af914f1a48a45","last_reissued_at":"2026-05-18T00:22:39.291334Z","signature_status":"signed_v1","first_computed_at":"2026-05-18T00:22:39.291334Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Backpropagation through the Void: Optimizing control variates for black-box gradient estimation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Dami Choi, David Duvenaud, Geoffrey Roeder, Will Grathwohl, Yuhuai Wu","submitted_at":"2017-10-31T21:58:13Z","abstract_excerpt":"Gradient-based optimization is the foundation of deep learning and reinforcement learning. Even when the mechanism being optimized is unknown or not differentiable, optimization using high-variance or biased gradient estimates is still often the best strategy. We introduce a general framework for learning low-variance, unbiased gradient estimators for black-box functions of random variables. Our method uses gradients of a neural network trained jointly with model parameters or policies, and is applicable in both discrete and continuous settings. We demonstrate this framework for training discr"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1711.00123","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":"1711.00123","created_at":"2026-05-18T00:22:39.291393+00:00"},{"alias_kind":"arxiv_version","alias_value":"1711.00123v3","created_at":"2026-05-18T00:22:39.291393+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1711.00123","created_at":"2026-05-18T00:22:39.291393+00:00"},{"alias_kind":"pith_short_12","alias_value":"5LR5GJA3WHNF","created_at":"2026-05-18T12:31:00.734936+00:00"},{"alias_kind":"pith_short_16","alias_value":"5LR5GJA3WHNFQTWW","created_at":"2026-05-18T12:31:00.734936+00:00"},{"alias_kind":"pith_short_8","alias_value":"5LR5GJA3","created_at":"2026-05-18T12:31:00.734936+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.03413","citing_title":"Learning to Theorize the World from Observation","ref_index":162,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/5LR5GJA3WHNFQTWWNTBY2A5JYP","json":"https://pith.science/pith/5LR5GJA3WHNFQTWWNTBY2A5JYP.json","graph_json":"https://pith.science/api/pith-number/5LR5GJA3WHNFQTWWNTBY2A5JYP/graph.json","events_json":"https://pith.science/api/pith-number/5LR5GJA3WHNFQTWWNTBY2A5JYP/events.json","paper":"https://pith.science/paper/5LR5GJA3"},"agent_actions":{"view_html":"https://pith.science/pith/5LR5GJA3WHNFQTWWNTBY2A5JYP","download_json":"https://pith.science/pith/5LR5GJA3WHNFQTWWNTBY2A5JYP.json","view_paper":"https://pith.science/paper/5LR5GJA3","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=1711.00123&json=true","fetch_graph":"https://pith.science/api/pith-number/5LR5GJA3WHNFQTWWNTBY2A5JYP/graph.json","fetch_events":"https://pith.science/api/pith-number/5LR5GJA3WHNFQTWWNTBY2A5JYP/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/5LR5GJA3WHNFQTWWNTBY2A5JYP/action/timestamp_anchor","attest_storage":"https://pith.science/pith/5LR5GJA3WHNFQTWWNTBY2A5JYP/action/storage_attestation","attest_author":"https://pith.science/pith/5LR5GJA3WHNFQTWWNTBY2A5JYP/action/author_attestation","sign_citation":"https://pith.science/pith/5LR5GJA3WHNFQTWWNTBY2A5JYP/action/citation_signature","submit_replication":"https://pith.science/pith/5LR5GJA3WHNFQTWWNTBY2A5JYP/action/replication_record"}},"created_at":"2026-05-18T00:22:39.291393+00:00","updated_at":"2026-05-18T00:22:39.291393+00:00"}