{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2018:VPBEQCPFPK5HGMBRA3JRQQHH46","short_pith_number":"pith:VPBEQCPF","schema_version":"1.0","canonical_sha256":"abc24809e57aba73303106d31840e7e796b7e69407c5e3b517f212b65e3bcb8e","source":{"kind":"arxiv","id":"1807.11143","version":2},"attestation_state":"computed","paper":{"title":"ARM: Augment-REINFORCE-Merge Gradient for Stochastic Binary Networks","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG","stat.CO","stat.ME"],"primary_cat":"stat.ML","authors_text":"Mingyuan Zhou, Mingzhang Yin","submitted_at":"2018-07-30T02:21:07Z","abstract_excerpt":"To backpropagate the gradients through stochastic binary layers, we propose the augment-REINFORCE-merge (ARM) estimator that is unbiased, exhibits low variance, and has low computational complexity. Exploiting variable augmentation, REINFORCE, and reparameterization, the ARM estimator achieves adaptive variance reduction for Monte Carlo integration by merging two expectations via common random numbers. The variance-reduction mechanism of the ARM estimator can also be attributed to either antithetic sampling in an augmented space, or the use of an optimal anti-symmetric \"self-control\" baseline "},"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":"1807.11143","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"stat.ML","submitted_at":"2018-07-30T02:21:07Z","cross_cats_sorted":["cs.LG","stat.CO","stat.ME"],"title_canon_sha256":"0c410bf2bcace4058d5ac05cb45ef3b408acab384005aac103b1f8ec893bfc3c","abstract_canon_sha256":"91351ed05a65075aed2b96f09bfdde9fcbd49339c13f4f10a9bf973cc8ffd422"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T00:03:12.299818Z","signature_b64":"dx+btZRjgqUsmsaC0RjJDd8CI5FOdgnT49CRDuOrRqJqoNXNV71MpLbT/618UfvV1mXBaxYQzaExdNAzkLl7BA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"abc24809e57aba73303106d31840e7e796b7e69407c5e3b517f212b65e3bcb8e","last_reissued_at":"2026-07-05T00:03:12.299169Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T00:03:12.299169Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"ARM: Augment-REINFORCE-Merge Gradient for Stochastic Binary Networks","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG","stat.CO","stat.ME"],"primary_cat":"stat.ML","authors_text":"Mingyuan Zhou, Mingzhang Yin","submitted_at":"2018-07-30T02:21:07Z","abstract_excerpt":"To backpropagate the gradients through stochastic binary layers, we propose the augment-REINFORCE-merge (ARM) estimator that is unbiased, exhibits low variance, and has low computational complexity. Exploiting variable augmentation, REINFORCE, and reparameterization, the ARM estimator achieves adaptive variance reduction for Monte Carlo integration by merging two expectations via common random numbers. The variance-reduction mechanism of the ARM estimator can also be attributed to either antithetic sampling in an augmented space, or the use of an optimal anti-symmetric \"self-control\" baseline "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1807.11143","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":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/1807.11143/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":"1807.11143","created_at":"2026-07-05T00:03:12.299231+00:00"},{"alias_kind":"arxiv_version","alias_value":"1807.11143v2","created_at":"2026-07-05T00:03:12.299231+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1807.11143","created_at":"2026-07-05T00:03:12.299231+00:00"},{"alias_kind":"pith_short_12","alias_value":"VPBEQCPFPK5H","created_at":"2026-07-05T00:03:12.299231+00:00"},{"alias_kind":"pith_short_16","alias_value":"VPBEQCPFPK5HGMBR","created_at":"2026-07-05T00:03:12.299231+00:00"},{"alias_kind":"pith_short_8","alias_value":"VPBEQCPF","created_at":"2026-07-05T00:03:12.299231+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2505.17962","citing_title":"A Principled Bayesian Framework for Training Binary and Spiking Neural Networks","ref_index":22,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/VPBEQCPFPK5HGMBRA3JRQQHH46","json":"https://pith.science/pith/VPBEQCPFPK5HGMBRA3JRQQHH46.json","graph_json":"https://pith.science/api/pith-number/VPBEQCPFPK5HGMBRA3JRQQHH46/graph.json","events_json":"https://pith.science/api/pith-number/VPBEQCPFPK5HGMBRA3JRQQHH46/events.json","paper":"https://pith.science/paper/VPBEQCPF"},"agent_actions":{"view_html":"https://pith.science/pith/VPBEQCPFPK5HGMBRA3JRQQHH46","download_json":"https://pith.science/pith/VPBEQCPFPK5HGMBRA3JRQQHH46.json","view_paper":"https://pith.science/paper/VPBEQCPF","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=1807.11143&json=true","fetch_graph":"https://pith.science/api/pith-number/VPBEQCPFPK5HGMBRA3JRQQHH46/graph.json","fetch_events":"https://pith.science/api/pith-number/VPBEQCPFPK5HGMBRA3JRQQHH46/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/VPBEQCPFPK5HGMBRA3JRQQHH46/action/timestamp_anchor","attest_storage":"https://pith.science/pith/VPBEQCPFPK5HGMBRA3JRQQHH46/action/storage_attestation","attest_author":"https://pith.science/pith/VPBEQCPFPK5HGMBRA3JRQQHH46/action/author_attestation","sign_citation":"https://pith.science/pith/VPBEQCPFPK5HGMBRA3JRQQHH46/action/citation_signature","submit_replication":"https://pith.science/pith/VPBEQCPFPK5HGMBRA3JRQQHH46/action/replication_record"}},"created_at":"2026-07-05T00:03:12.299231+00:00","updated_at":"2026-07-05T00:03:12.299231+00:00"}