{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2022:MHYEWR5Y35IGYMSFLPOWWWI5ID","short_pith_number":"pith:MHYEWR5Y","canonical_record":{"source":{"id":"2202.11818","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2022-02-23T23:00:40Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"aeb018d062ba36d2cf45786f0a897aa5fc6234a51f500e6cd8e542d312e8afe0","abstract_canon_sha256":"23c04596bdd5fdc966307ae0bb6b50db14d1c8b58e85643bc5443e1612f3c60c"},"schema_version":"1.0"},"canonical_sha256":"61f04b47b8df506c32455bdd6b591d40f9d3368598b34a56b1a828f34bde2fcb","source":{"kind":"arxiv","id":"2202.11818","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2202.11818","created_at":"2026-07-05T03:59:46Z"},{"alias_kind":"arxiv_version","alias_value":"2202.11818v1","created_at":"2026-07-05T03:59:46Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2202.11818","created_at":"2026-07-05T03:59:46Z"},{"alias_kind":"pith_short_12","alias_value":"MHYEWR5Y35IG","created_at":"2026-07-05T03:59:46Z"},{"alias_kind":"pith_short_16","alias_value":"MHYEWR5Y35IGYMSF","created_at":"2026-07-05T03:59:46Z"},{"alias_kind":"pith_short_8","alias_value":"MHYEWR5Y","created_at":"2026-07-05T03:59:46Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2022:MHYEWR5Y35IGYMSFLPOWWWI5ID","target":"record","payload":{"canonical_record":{"source":{"id":"2202.11818","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2022-02-23T23:00:40Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"aeb018d062ba36d2cf45786f0a897aa5fc6234a51f500e6cd8e542d312e8afe0","abstract_canon_sha256":"23c04596bdd5fdc966307ae0bb6b50db14d1c8b58e85643bc5443e1612f3c60c"},"schema_version":"1.0"},"canonical_sha256":"61f04b47b8df506c32455bdd6b591d40f9d3368598b34a56b1a828f34bde2fcb","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T03:59:46.440839Z","signature_b64":"hJx3BPTibTuX4QI+8YjnkyFY3+wphy/suF8yz9VstI/TDc0eCGbeH46w1HXQsgTpSWvh8K9LGsxtToeIk7QKBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"61f04b47b8df506c32455bdd6b591d40f9d3368598b34a56b1a828f34bde2fcb","last_reissued_at":"2026-07-05T03:59:46.440424Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T03:59:46.440424Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2202.11818","source_version":1,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T03:59:46Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"eA+E71q4X14L9bONpz7BuwhT43M/DNT8eWlRVsRbTbRWWDOsyW/+ihsEH252yLD293hclwnlHw2SVXnYj8mpDQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-08T04:42:48.948852Z"},"content_sha256":"b22023f86f4ea758feedf629f452170ff51549bf20efa44b5688706e014d0d43","schema_version":"1.0","event_id":"sha256:b22023f86f4ea758feedf629f452170ff51549bf20efa44b5688706e014d0d43"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2022:MHYEWR5Y35IGYMSFLPOWWWI5ID","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Consistent Dropout for Policy Gradient Reinforcement Learning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Matthew Hausknecht, Nolan Wagener","submitted_at":"2022-02-23T23:00:40Z","abstract_excerpt":"Dropout has long been a staple of supervised learning, but is rarely used in reinforcement learning. We analyze why naive application of dropout is problematic for policy-gradient learning algorithms and introduce consistent dropout, a simple technique to address this instability. We demonstrate consistent dropout enables stable training with A2C and PPO in both continuous and discrete action environments across a wide range of dropout probabilities. Finally, we show that consistent dropout enables the online training of complex architectures such as GPT without needing to disable the model's "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2202.11818","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/2202.11818/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"},"verdict_id":null},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T03:59:46Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"h8dNvJq2oyW3LgjgZRFgos4qQMYflUS3TOnY0n9lBJItZMp0hhtaUhDOgihkYFP41bwxJc4Ab1iOdv1laKiCBA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-08T04:42:48.949428Z"},"content_sha256":"c6bf8b08f3d7db8d8812a699c6440bb6de5e310c2fa37b74ee525a597202a5d2","schema_version":"1.0","event_id":"sha256:c6bf8b08f3d7db8d8812a699c6440bb6de5e310c2fa37b74ee525a597202a5d2"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/MHYEWR5Y35IGYMSFLPOWWWI5ID/bundle.json","state_url":"https://pith.science/pith/MHYEWR5Y35IGYMSFLPOWWWI5ID/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/MHYEWR5Y35IGYMSFLPOWWWI5ID/bundle.json","status":"primary"}],"public_keys":[{"key_id":"pith-v1-2026-05","algorithm":"ed25519","format":"raw","public_key_b64":"stVStoiQhXFxp4s2pdzPNoqVNBMojDU/fJ2db5S3CbM=","public_key_hex":"b2d552b68890857171a78b36a5dccf368a953413288c353f7c9d9d6f94b709b3","fingerprint_sha256_b32_first128bits":"RVFV5Z2OI2J3ZUO7ERDEBCYNKS","fingerprint_sha256_hex":"8d4b5ee74e4693bcd1df2446408b0d54","rotates_at":null,"url":"https://pith.science/pith-signing-key.json","notes":"Pith uses this Ed25519 key to sign canonical record SHA-256 digests. Verify with: ed25519_verify(public_key, message=canonical_sha256_bytes, signature=base64decode(signature_b64))."}],"merge_version":"pith-open-graph-merge-v1","built_at":"2026-08-08T04:42:48Z","links":{"resolver":"https://pith.science/pith/MHYEWR5Y35IGYMSFLPOWWWI5ID","bundle":"https://pith.science/pith/MHYEWR5Y35IGYMSFLPOWWWI5ID/bundle.json","state":"https://pith.science/pith/MHYEWR5Y35IGYMSFLPOWWWI5ID/state.json","well_known_bundle":"https://pith.science/.well-known/pith/MHYEWR5Y35IGYMSFLPOWWWI5ID/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2022:MHYEWR5Y35IGYMSFLPOWWWI5ID","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"23c04596bdd5fdc966307ae0bb6b50db14d1c8b58e85643bc5443e1612f3c60c","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2022-02-23T23:00:40Z","title_canon_sha256":"aeb018d062ba36d2cf45786f0a897aa5fc6234a51f500e6cd8e542d312e8afe0"},"schema_version":"1.0","source":{"id":"2202.11818","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2202.11818","created_at":"2026-07-05T03:59:46Z"},{"alias_kind":"arxiv_version","alias_value":"2202.11818v1","created_at":"2026-07-05T03:59:46Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2202.11818","created_at":"2026-07-05T03:59:46Z"},{"alias_kind":"pith_short_12","alias_value":"MHYEWR5Y35IG","created_at":"2026-07-05T03:59:46Z"},{"alias_kind":"pith_short_16","alias_value":"MHYEWR5Y35IGYMSF","created_at":"2026-07-05T03:59:46Z"},{"alias_kind":"pith_short_8","alias_value":"MHYEWR5Y","created_at":"2026-07-05T03:59:46Z"}],"graph_snapshots":[{"event_id":"sha256:c6bf8b08f3d7db8d8812a699c6440bb6de5e310c2fa37b74ee525a597202a5d2","target":"graph","created_at":"2026-07-05T03:59:46Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2202.11818/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Dropout has long been a staple of supervised learning, but is rarely used in reinforcement learning. We analyze why naive application of dropout is problematic for policy-gradient learning algorithms and introduce consistent dropout, a simple technique to address this instability. We demonstrate consistent dropout enables stable training with A2C and PPO in both continuous and discrete action environments across a wide range of dropout probabilities. Finally, we show that consistent dropout enables the online training of complex architectures such as GPT without needing to disable the model's ","authors_text":"Matthew Hausknecht, Nolan Wagener","cross_cats":["cs.AI"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2022-02-23T23:00:40Z","title":"Consistent Dropout for Policy Gradient Reinforcement Learning"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2202.11818","kind":"arxiv","version":1},"verdict":{"created_at":null,"id":null,"model_set":{},"one_line_summary":"","pipeline_version":null,"pith_extraction_headline":"","strongest_claim":"","weakest_assumption":""}},"verdict_id":null}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:b22023f86f4ea758feedf629f452170ff51549bf20efa44b5688706e014d0d43","target":"record","created_at":"2026-07-05T03:59:46Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"23c04596bdd5fdc966307ae0bb6b50db14d1c8b58e85643bc5443e1612f3c60c","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2022-02-23T23:00:40Z","title_canon_sha256":"aeb018d062ba36d2cf45786f0a897aa5fc6234a51f500e6cd8e542d312e8afe0"},"schema_version":"1.0","source":{"id":"2202.11818","kind":"arxiv","version":1}},"canonical_sha256":"61f04b47b8df506c32455bdd6b591d40f9d3368598b34a56b1a828f34bde2fcb","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"61f04b47b8df506c32455bdd6b591d40f9d3368598b34a56b1a828f34bde2fcb","first_computed_at":"2026-07-05T03:59:46.440424Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T03:59:46.440424Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"hJx3BPTibTuX4QI+8YjnkyFY3+wphy/suF8yz9VstI/TDc0eCGbeH46w1HXQsgTpSWvh8K9LGsxtToeIk7QKBg==","signature_status":"signed_v1","signed_at":"2026-07-05T03:59:46.440839Z","signed_message":"canonical_sha256_bytes"},"source_id":"2202.11818","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:b22023f86f4ea758feedf629f452170ff51549bf20efa44b5688706e014d0d43","sha256:c6bf8b08f3d7db8d8812a699c6440bb6de5e310c2fa37b74ee525a597202a5d2"],"state_sha256":"f0b74d720c13e3889abe3a78f69384279f3dc25ae57dd000b1604879e4dcd4df"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"3qfBzZ8q0P2nQ+D/VnsHF+iCvRZ5h7eeRDJ0sjKIPnuN0gsbxZ8lYbLrTS7ZXiBjEY3qGtIlmkeVi1qWJkvvDQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-08T04:42:48.953443Z","bundle_sha256":"9fe2577695afa2af24f69be7a00e6196a39efa49bb9ee81c06e36f52d57e15e9"}}