{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2018:RGJQSOASTZRVN7MORTM5TSH7P4","short_pith_number":"pith:RGJQSOAS","canonical_record":{"source":{"id":"1805.07805","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2018-05-20T17:47:03Z","cross_cats_sorted":["cs.AI","stat.ML"],"title_canon_sha256":"1d03d38cb1eb4febc569fad8ba33445efa8deb535aca657893432ddbbe3e8fc2","abstract_canon_sha256":"915d730a56175d8a05baa7d5b498d68d1bd0f3a0f47430ce9c581741ebe00025"},"schema_version":"1.0"},"canonical_sha256":"89930938129e6356fd8e8cd9d9c8ff7f35a5d6d61e55bfa27e5449edf236e8af","source":{"kind":"arxiv","id":"1805.07805","version":3},"source_aliases":[{"alias_kind":"arxiv","alias_value":"1805.07805","created_at":"2026-05-17T23:40:56Z"},{"alias_kind":"arxiv_version","alias_value":"1805.07805v3","created_at":"2026-05-17T23:40:56Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1805.07805","created_at":"2026-05-17T23:40:56Z"},{"alias_kind":"pith_short_12","alias_value":"RGJQSOASTZRV","created_at":"2026-05-18T12:32:50Z"},{"alias_kind":"pith_short_16","alias_value":"RGJQSOASTZRVN7MO","created_at":"2026-05-18T12:32:50Z"},{"alias_kind":"pith_short_8","alias_value":"RGJQSOAS","created_at":"2026-05-18T12:32:50Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2018:RGJQSOASTZRVN7MORTM5TSH7P4","target":"record","payload":{"canonical_record":{"source":{"id":"1805.07805","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2018-05-20T17:47:03Z","cross_cats_sorted":["cs.AI","stat.ML"],"title_canon_sha256":"1d03d38cb1eb4febc569fad8ba33445efa8deb535aca657893432ddbbe3e8fc2","abstract_canon_sha256":"915d730a56175d8a05baa7d5b498d68d1bd0f3a0f47430ce9c581741ebe00025"},"schema_version":"1.0"},"canonical_sha256":"89930938129e6356fd8e8cd9d9c8ff7f35a5d6d61e55bfa27e5449edf236e8af","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-17T23:40:56.060106Z","signature_b64":"+SpGw4UyEbTN8AqJFpKL6rc0b8eQEJ7PObq4G0KGvK5pUoRTVXRC47mD8vwicN6yUOXV1RSJFnIX/0HjlrdKBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"89930938129e6356fd8e8cd9d9c8ff7f35a5d6d61e55bfa27e5449edf236e8af","last_reissued_at":"2026-05-17T23:40:56.059556Z","signature_status":"signed_v1","first_computed_at":"2026-05-17T23:40:56.059556Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"1805.07805","source_version":3,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-17T23:40:56Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"q8pby/vsD1qz6JzOuET+nkUMAxwfWHpK18QCpcwBb2rrHJmNrEa4AYQ28OORtqdId0aGyipRYP/MeipQdJZyCg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-11T06:00:54.747231Z"},"content_sha256":"557cc1c6072e03de0acfc01670ef86acfaecc857792e89d7b990c92dd4eb220a","schema_version":"1.0","event_id":"sha256:557cc1c6072e03de0acfc01670ef86acfaecc857792e89d7b990c92dd4eb220a"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2018:RGJQSOASTZRVN7MORTM5TSH7P4","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Constrained Policy Improvement for Safe and Efficient Reinforcement Learning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","stat.ML"],"primary_cat":"cs.LG","authors_text":"Aviv Tamar, Elad Sarafian, Sarit Kraus","submitted_at":"2018-05-20T17:47:03Z","abstract_excerpt":"We propose a policy improvement algorithm for Reinforcement Learning (RL) which is called Rerouted Behavior Improvement (RBI). RBI is designed to take into account the evaluation errors of the Q-function. Such errors are common in RL when learning the $Q$-value from finite past experience data. Greedy policies or even constrained policy optimization algorithms which ignore these errors may suffer from an improvement penalty (i.e. a negative policy improvement). To minimize the improvement penalty, the RBI idea is to attenuate rapid policy changes of low probability actions which were less freq"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1805.07805","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"},"verdict_id":null},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-17T23:40:56Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"oBS3a9nQFZjoxJAto+nELXHm32ZPrcVgYnjfSs8qKUkSZXq1Pyauiv/xqgH97HDJsWlwsMYGdwCpl/r7QqHwBQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-11T06:00:54.747998Z"},"content_sha256":"3788924265a2c4a5f63ab2c0f4c61d939c0915e75bb575aae4bd00eb118226fd","schema_version":"1.0","event_id":"sha256:3788924265a2c4a5f63ab2c0f4c61d939c0915e75bb575aae4bd00eb118226fd"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/RGJQSOASTZRVN7MORTM5TSH7P4/bundle.json","state_url":"https://pith.science/pith/RGJQSOASTZRVN7MORTM5TSH7P4/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/RGJQSOASTZRVN7MORTM5TSH7P4/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":"2026-06-11T06:00:54Z","links":{"resolver":"https://pith.science/pith/RGJQSOASTZRVN7MORTM5TSH7P4","bundle":"https://pith.science/pith/RGJQSOASTZRVN7MORTM5TSH7P4/bundle.json","state":"https://pith.science/pith/RGJQSOASTZRVN7MORTM5TSH7P4/state.json","well_known_bundle":"https://pith.science/.well-known/pith/RGJQSOASTZRVN7MORTM5TSH7P4/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2018:RGJQSOASTZRVN7MORTM5TSH7P4","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":"915d730a56175d8a05baa7d5b498d68d1bd0f3a0f47430ce9c581741ebe00025","cross_cats_sorted":["cs.AI","stat.ML"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2018-05-20T17:47:03Z","title_canon_sha256":"1d03d38cb1eb4febc569fad8ba33445efa8deb535aca657893432ddbbe3e8fc2"},"schema_version":"1.0","source":{"id":"1805.07805","kind":"arxiv","version":3}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"1805.07805","created_at":"2026-05-17T23:40:56Z"},{"alias_kind":"arxiv_version","alias_value":"1805.07805v3","created_at":"2026-05-17T23:40:56Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1805.07805","created_at":"2026-05-17T23:40:56Z"},{"alias_kind":"pith_short_12","alias_value":"RGJQSOASTZRV","created_at":"2026-05-18T12:32:50Z"},{"alias_kind":"pith_short_16","alias_value":"RGJQSOASTZRVN7MO","created_at":"2026-05-18T12:32:50Z"},{"alias_kind":"pith_short_8","alias_value":"RGJQSOAS","created_at":"2026-05-18T12:32:50Z"}],"graph_snapshots":[{"event_id":"sha256:3788924265a2c4a5f63ab2c0f4c61d939c0915e75bb575aae4bd00eb118226fd","target":"graph","created_at":"2026-05-17T23:40:56Z","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"},"paper":{"abstract_excerpt":"We propose a policy improvement algorithm for Reinforcement Learning (RL) which is called Rerouted Behavior Improvement (RBI). RBI is designed to take into account the evaluation errors of the Q-function. Such errors are common in RL when learning the $Q$-value from finite past experience data. Greedy policies or even constrained policy optimization algorithms which ignore these errors may suffer from an improvement penalty (i.e. a negative policy improvement). To minimize the improvement penalty, the RBI idea is to attenuate rapid policy changes of low probability actions which were less freq","authors_text":"Aviv Tamar, Elad Sarafian, Sarit Kraus","cross_cats":["cs.AI","stat.ML"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2018-05-20T17:47:03Z","title":"Constrained Policy Improvement for Safe and Efficient Reinforcement Learning"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1805.07805","kind":"arxiv","version":3},"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":"sha256:557cc1c6072e03de0acfc01670ef86acfaecc857792e89d7b990c92dd4eb220a","target":"record","created_at":"2026-05-17T23:40:56Z","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":"915d730a56175d8a05baa7d5b498d68d1bd0f3a0f47430ce9c581741ebe00025","cross_cats_sorted":["cs.AI","stat.ML"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2018-05-20T17:47:03Z","title_canon_sha256":"1d03d38cb1eb4febc569fad8ba33445efa8deb535aca657893432ddbbe3e8fc2"},"schema_version":"1.0","source":{"id":"1805.07805","kind":"arxiv","version":3}},"canonical_sha256":"89930938129e6356fd8e8cd9d9c8ff7f35a5d6d61e55bfa27e5449edf236e8af","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"89930938129e6356fd8e8cd9d9c8ff7f35a5d6d61e55bfa27e5449edf236e8af","first_computed_at":"2026-05-17T23:40:56.059556Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-17T23:40:56.059556Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"+SpGw4UyEbTN8AqJFpKL6rc0b8eQEJ7PObq4G0KGvK5pUoRTVXRC47mD8vwicN6yUOXV1RSJFnIX/0HjlrdKBg==","signature_status":"signed_v1","signed_at":"2026-05-17T23:40:56.060106Z","signed_message":"canonical_sha256_bytes"},"source_id":"1805.07805","source_kind":"arxiv","source_version":3}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:557cc1c6072e03de0acfc01670ef86acfaecc857792e89d7b990c92dd4eb220a","sha256:3788924265a2c4a5f63ab2c0f4c61d939c0915e75bb575aae4bd00eb118226fd"],"state_sha256":"e5eda2e8f2d7d7f980575edaba4e8aaa83b2bd55b0c2ab482979664291f36dde"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"iCIh8keGB4X7vSI5CVVGTWsaU3qeOuuyB2rUZbXoQg9TKlDVri4lKJjweZJP8c1AWMHiEiXV3h3dPhFAeXCSCg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-06-11T06:00:54.752847Z","bundle_sha256":"f16a967340a616fbcc951bdb993801019f63dee48daadba94a5f6a437c44e559"}}