{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:T4MMAUXIQFSLZRUXZKKIG7K6YR","short_pith_number":"pith:T4MMAUXI","canonical_record":{"source":{"id":"2607.08925","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-07-09T20:41:45Z","cross_cats_sorted":[],"title_canon_sha256":"d69c037e15e260b13993dd23d0a9d4e29de85d6aede1a04cbbcece6b1cc5fbad","abstract_canon_sha256":"c1a0c0f1b7ed8123312f7826bb338297534e748786407f955eef44f9bbe351d7"},"schema_version":"1.0"},"canonical_sha256":"9f18c052e88164bcc697ca94837d5ec457accd2b7a854a19a199d254de75bdb4","source":{"kind":"arxiv","id":"2607.08925","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2607.08925","created_at":"2026-07-13T00:17:27Z"},{"alias_kind":"arxiv_version","alias_value":"2607.08925v1","created_at":"2026-07-13T00:17:27Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2607.08925","created_at":"2026-07-13T00:17:27Z"},{"alias_kind":"pith_short_12","alias_value":"T4MMAUXIQFSL","created_at":"2026-07-13T00:17:27Z"},{"alias_kind":"pith_short_16","alias_value":"T4MMAUXIQFSLZRUX","created_at":"2026-07-13T00:17:27Z"},{"alias_kind":"pith_short_8","alias_value":"T4MMAUXI","created_at":"2026-07-13T00:17:27Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:T4MMAUXIQFSLZRUXZKKIG7K6YR","target":"record","payload":{"canonical_record":{"source":{"id":"2607.08925","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-07-09T20:41:45Z","cross_cats_sorted":[],"title_canon_sha256":"d69c037e15e260b13993dd23d0a9d4e29de85d6aede1a04cbbcece6b1cc5fbad","abstract_canon_sha256":"c1a0c0f1b7ed8123312f7826bb338297534e748786407f955eef44f9bbe351d7"},"schema_version":"1.0"},"canonical_sha256":"9f18c052e88164bcc697ca94837d5ec457accd2b7a854a19a199d254de75bdb4","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-13T00:17:27.359965Z","signature_b64":"XKq9i8r2OCN7Rv7LjiuLcsBcg2aBeVizcBCxlqq2J8vzFC5s8IyOVKVtWZSCpYZYBLSWGW/cRySGXphSrtkBCA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"9f18c052e88164bcc697ca94837d5ec457accd2b7a854a19a199d254de75bdb4","last_reissued_at":"2026-07-13T00:17:27.358969Z","signature_status":"signed_v1","first_computed_at":"2026-07-13T00:17:27.358969Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2607.08925","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":"2026-07-13T00:17:27Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"ktR3l1yvTsOIHbzHmW/1rLLdWL7B5VYN2jryUEU8HtrIilwt+2H5Di9Hyr08J9f5UCSfgGT/c8jUtkPeO9SIDg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-08T07:44:58.232327Z"},"content_sha256":"a2085fbe2000825634f6ad83e7e6d23e704c2d45bae3a4e0c9c3638d3649dda8","schema_version":"1.0","event_id":"sha256:a2085fbe2000825634f6ad83e7e6d23e704c2d45bae3a4e0c9c3638d3649dda8"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:T4MMAUXIQFSLZRUXZKKIG7K6YR","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"SafeExplorer: An Unbiased Policy Gradient for Reinforcement Learning with Recovery Interventions","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Elham Daneshmand, Glen Berseth, Hsiu-Chin Lin, Majid Khadiv","submitted_at":"2026-07-09T20:41:45Z","abstract_excerpt":"Training reinforcement-learning agents directly on physical robots makes every fall costly, since a fall can damage the platform and cannot be undone like a simulator reset; the goal is therefore to minimize falls during training rather than trade them off against return, as constrained Markov decision process (MDP) formulations do. A standard mitigation hands control to a separate recovery policy whenever the agent leaves a designer-specified safe region (a subset of state space it should stay within), but the resulting mixed-policy rollouts silently bias every on-policy update, and the impor"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2607.08925","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/2607.08925/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":"2026-07-13T00:17:27Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"k6T2+N48asbdbjcusdXbB4P136y7snFJVBwhhgfZQLBBWL3T4nd7jimvvvXYOdq3imAVd+xyry2BtzBu2EecAw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-08T07:44:58.232823Z"},"content_sha256":"d6fc6afba2830e201afc9cd202068b17a31b87e0ebc239da27321ad052b227e6","schema_version":"1.0","event_id":"sha256:d6fc6afba2830e201afc9cd202068b17a31b87e0ebc239da27321ad052b227e6"},{"event_type":"integrity_finding","subject_pith_number":"pith:2026:T4MMAUXIQFSLZRUXZKKIG7K6YR","target":"integrity","payload":{"note":"DOI in the printed bibliography is fragmented by whitespace or line breaks. A longer candidate (10.1109/Humanoids58906.2024.10769799.Alex) was visible in the surrounding text but could not be confirmed against doi.org as printed.","snippet":"doi: 10.1109/Humanoids58906. 2024.10769799. Alex Ray, Joshua Achiam, and Dario Amodei. Benchmarking safe exploration in deep reinforcement learning. Technical report, OpenAI,","arxiv_id":"2607.08925","detector":"doi_compliance","evidence":{"ref_index":12,"verdict_class":"incontrovertible","resolved_title":null,"printed_excerpt":"doi: 10.1109/Humanoids58906. 2024.10769799. Alex Ray, Joshua Achiam, and Dario Amodei. Benchmarking safe exploration in deep reinforcement learning. Technical report, OpenAI,","reconstructed_doi":"10.1109/Humanoids58906.2024.10769799.Alex"},"severity":"advisory","ref_index":12,"audited_at":"2026-07-13T05:55:24.490744Z","event_type":"pith.integrity.v1","detected_doi":"10.1109/Humanoids58906.2024.10769799.Alex","detector_url":"https://pith.science/pith-integrity-protocol#doi_compliance","external_url":null,"finding_type":"recoverable_identifier","evidence_hash":"8c1791ed553032111bb5f6c44c8b6b7ea5a04d6d823f1e12ce06058d74d73201","paper_version":1,"verdict_class":"incontrovertible","resolved_title":null,"detector_version":"1.0.0","detected_arxiv_id":null,"integrity_event_id":12984,"payload_sha256":"82083dc95781839a87c1b304959248266b2210647e6241ba5b10455ab98c02ea","signature_b64":"zBTm+2AVaz0PDozgKRPlHlLXUVnYRHk8Bez5D/wSJT+KqJ79K2uvAMyc4zb7mFMv+5dpsPmAOs9h8iyaZ/WXCw==","signing_key_id":"pith-v1-2026-05"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-13T05:58:53Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"8Zv5l5/o9Ma3RVp+Kml7IONBNaX701lpkxPqWuNZRb66wrfeFCzPjxDzZnycN7tOmVcH3hoybJMn0e5MgiArCA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-08T07:44:58.236264Z"},"content_sha256":"d00488c7c7ffe72dec4958f1a2b8d524151346f36bed4c8e76e9001fe04ad4ba","schema_version":"1.0","event_id":"sha256:d00488c7c7ffe72dec4958f1a2b8d524151346f36bed4c8e76e9001fe04ad4ba"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/T4MMAUXIQFSLZRUXZKKIG7K6YR/bundle.json","state_url":"https://pith.science/pith/T4MMAUXIQFSLZRUXZKKIG7K6YR/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/T4MMAUXIQFSLZRUXZKKIG7K6YR/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":"2026-08-08T07:44:58Z","links":{"resolver":"https://pith.science/pith/T4MMAUXIQFSLZRUXZKKIG7K6YR","bundle":"https://pith.science/pith/T4MMAUXIQFSLZRUXZKKIG7K6YR/bundle.json","state":"https://pith.science/pith/T4MMAUXIQFSLZRUXZKKIG7K6YR/state.json","well_known_bundle":"https://pith.science/.well-known/pith/T4MMAUXIQFSLZRUXZKKIG7K6YR/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:T4MMAUXIQFSLZRUXZKKIG7K6YR","merge_version":"pith-open-graph-merge-v1","event_count":3,"valid_event_count":3,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"c1a0c0f1b7ed8123312f7826bb338297534e748786407f955eef44f9bbe351d7","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-07-09T20:41:45Z","title_canon_sha256":"d69c037e15e260b13993dd23d0a9d4e29de85d6aede1a04cbbcece6b1cc5fbad"},"schema_version":"1.0","source":{"id":"2607.08925","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2607.08925","created_at":"2026-07-13T00:17:27Z"},{"alias_kind":"arxiv_version","alias_value":"2607.08925v1","created_at":"2026-07-13T00:17:27Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2607.08925","created_at":"2026-07-13T00:17:27Z"},{"alias_kind":"pith_short_12","alias_value":"T4MMAUXIQFSL","created_at":"2026-07-13T00:17:27Z"},{"alias_kind":"pith_short_16","alias_value":"T4MMAUXIQFSLZRUX","created_at":"2026-07-13T00:17:27Z"},{"alias_kind":"pith_short_8","alias_value":"T4MMAUXI","created_at":"2026-07-13T00:17:27Z"}],"graph_snapshots":[{"event_id":"sha256:d6fc6afba2830e201afc9cd202068b17a31b87e0ebc239da27321ad052b227e6","target":"graph","created_at":"2026-07-13T00:17:27Z","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/2607.08925/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Training reinforcement-learning agents directly on physical robots makes every fall costly, since a fall can damage the platform and cannot be undone like a simulator reset; the goal is therefore to minimize falls during training rather than trade them off against return, as constrained Markov decision process (MDP) formulations do. A standard mitigation hands control to a separate recovery policy whenever the agent leaves a designer-specified safe region (a subset of state space it should stay within), but the resulting mixed-policy rollouts silently bias every on-policy update, and the impor","authors_text":"Elham Daneshmand, Glen Berseth, Hsiu-Chin Lin, Majid Khadiv","cross_cats":[],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-07-09T20:41:45Z","title":"SafeExplorer: An Unbiased Policy Gradient for Reinforcement Learning with Recovery Interventions"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2607.08925","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:a2085fbe2000825634f6ad83e7e6d23e704c2d45bae3a4e0c9c3638d3649dda8","target":"record","created_at":"2026-07-13T00:17:27Z","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":"c1a0c0f1b7ed8123312f7826bb338297534e748786407f955eef44f9bbe351d7","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-07-09T20:41:45Z","title_canon_sha256":"d69c037e15e260b13993dd23d0a9d4e29de85d6aede1a04cbbcece6b1cc5fbad"},"schema_version":"1.0","source":{"id":"2607.08925","kind":"arxiv","version":1}},"canonical_sha256":"9f18c052e88164bcc697ca94837d5ec457accd2b7a854a19a199d254de75bdb4","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"9f18c052e88164bcc697ca94837d5ec457accd2b7a854a19a199d254de75bdb4","first_computed_at":"2026-07-13T00:17:27.358969Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-13T00:17:27.358969Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"XKq9i8r2OCN7Rv7LjiuLcsBcg2aBeVizcBCxlqq2J8vzFC5s8IyOVKVtWZSCpYZYBLSWGW/cRySGXphSrtkBCA==","signature_status":"signed_v1","signed_at":"2026-07-13T00:17:27.359965Z","signed_message":"canonical_sha256_bytes"},"source_id":"2607.08925","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:a2085fbe2000825634f6ad83e7e6d23e704c2d45bae3a4e0c9c3638d3649dda8","sha256:d6fc6afba2830e201afc9cd202068b17a31b87e0ebc239da27321ad052b227e6","sha256:d00488c7c7ffe72dec4958f1a2b8d524151346f36bed4c8e76e9001fe04ad4ba"],"state_sha256":"4efc6e2099645ab42952001b0563460a525c7beb44e02d76f4f7fa5d3aefa51f"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"cYBLi6JyO6zNNXZNpc0iD4NHnd7pmEcD37RytBa8HUVtW9XI2FSqkPkcd7tiOmcD/D4nE0B57gl66qCVQxPFDw==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-08T07:44:58.238569Z","bundle_sha256":"be5e7fd8cac8caa3adc91b7a36de51df8eb86a3bde50e72b8475d6fcfbfe8240"}}