{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:FUPXWAFVSDX4B7X43TPRURDQYF","short_pith_number":"pith:FUPXWAFV","schema_version":"1.0","canonical_sha256":"2d1f7b00b590efc0fefcdcdf1a4470c141713285a41032fab22d24c2cb2458e4","source":{"kind":"arxiv","id":"2112.13835","version":1},"attestation_state":"computed","paper":{"title":"Unbiased Gradient Estimation in Unrolled Computation Graphs with Persistent Evolution Strategies","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["stat.ML"],"primary_cat":"cs.LG","authors_text":"Jascha Sohl-Dickstein, Luke Metz, Paul Vicol","submitted_at":"2021-12-27T18:54:36Z","abstract_excerpt":"Unrolled computation graphs arise in many scenarios, including training RNNs, tuning hyperparameters through unrolled optimization, and training learned optimizers. Current approaches to optimizing parameters in such computation graphs suffer from high variance gradients, bias, slow updates, or large memory usage. We introduce a method called Persistent Evolution Strategies (PES), which divides the computation graph into a series of truncated unrolls, and performs an evolution strategies-based update step after each unroll. PES eliminates bias from these truncations by accumulating correction "},"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":"2112.13835","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2021-12-27T18:54:36Z","cross_cats_sorted":["stat.ML"],"title_canon_sha256":"45489c6a153fec1c592682f2691e995048c151709d655144249edf4f414c1f36","abstract_canon_sha256":"6c942392fc581014fce0af61d25e0660b5a8b024b48f77a0ad98b384e8c2d86f"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T03:43:57.459077Z","signature_b64":"JAXGqe0WU/53PrpIBEImi1pOvSMpZhqimRkYA0Ub1mXs0HlUJRxTsTA7SXu592SRjS4ZEM/7kj7nR92EpvGxAQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"2d1f7b00b590efc0fefcdcdf1a4470c141713285a41032fab22d24c2cb2458e4","last_reissued_at":"2026-07-05T03:43:57.458604Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T03:43:57.458604Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Unbiased Gradient Estimation in Unrolled Computation Graphs with Persistent Evolution Strategies","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["stat.ML"],"primary_cat":"cs.LG","authors_text":"Jascha Sohl-Dickstein, Luke Metz, Paul Vicol","submitted_at":"2021-12-27T18:54:36Z","abstract_excerpt":"Unrolled computation graphs arise in many scenarios, including training RNNs, tuning hyperparameters through unrolled optimization, and training learned optimizers. Current approaches to optimizing parameters in such computation graphs suffer from high variance gradients, bias, slow updates, or large memory usage. We introduce a method called Persistent Evolution Strategies (PES), which divides the computation graph into a series of truncated unrolls, and performs an evolution strategies-based update step after each unroll. PES eliminates bias from these truncations by accumulating correction "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2112.13835","kind":"arxiv","version":1},"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/2112.13835/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":"2112.13835","created_at":"2026-07-05T03:43:57.458672+00:00"},{"alias_kind":"arxiv_version","alias_value":"2112.13835v1","created_at":"2026-07-05T03:43:57.458672+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2112.13835","created_at":"2026-07-05T03:43:57.458672+00:00"},{"alias_kind":"pith_short_12","alias_value":"FUPXWAFVSDX4","created_at":"2026-07-05T03:43:57.458672+00:00"},{"alias_kind":"pith_short_16","alias_value":"FUPXWAFVSDX4B7X4","created_at":"2026-07-05T03:43:57.458672+00:00"},{"alias_kind":"pith_short_8","alias_value":"FUPXWAFV","created_at":"2026-07-05T03:43:57.458672+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/FUPXWAFVSDX4B7X43TPRURDQYF","json":"https://pith.science/pith/FUPXWAFVSDX4B7X43TPRURDQYF.json","graph_json":"https://pith.science/api/pith-number/FUPXWAFVSDX4B7X43TPRURDQYF/graph.json","events_json":"https://pith.science/api/pith-number/FUPXWAFVSDX4B7X43TPRURDQYF/events.json","paper":"https://pith.science/paper/FUPXWAFV"},"agent_actions":{"view_html":"https://pith.science/pith/FUPXWAFVSDX4B7X43TPRURDQYF","download_json":"https://pith.science/pith/FUPXWAFVSDX4B7X43TPRURDQYF.json","view_paper":"https://pith.science/paper/FUPXWAFV","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2112.13835&json=true","fetch_graph":"https://pith.science/api/pith-number/FUPXWAFVSDX4B7X43TPRURDQYF/graph.json","fetch_events":"https://pith.science/api/pith-number/FUPXWAFVSDX4B7X43TPRURDQYF/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/FUPXWAFVSDX4B7X43TPRURDQYF/action/timestamp_anchor","attest_storage":"https://pith.science/pith/FUPXWAFVSDX4B7X43TPRURDQYF/action/storage_attestation","attest_author":"https://pith.science/pith/FUPXWAFVSDX4B7X43TPRURDQYF/action/author_attestation","sign_citation":"https://pith.science/pith/FUPXWAFVSDX4B7X43TPRURDQYF/action/citation_signature","submit_replication":"https://pith.science/pith/FUPXWAFVSDX4B7X43TPRURDQYF/action/replication_record"}},"created_at":"2026-07-05T03:43:57.458672+00:00","updated_at":"2026-07-05T03:43:57.458672+00:00"}