{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:DIUNESJXV3LCVKC5E2ABI2CX6J","short_pith_number":"pith:DIUNESJX","canonical_record":{"source":{"id":"2608.10204","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-08-10T20:20:01Z","cross_cats_sorted":[],"title_canon_sha256":"cca53521182367838dbadcb81c688b0e7581c5448163c4ffb25bcb5fb25f631b","abstract_canon_sha256":"2167452e6d880f4d503208b54e4b8fe38341aafeefb42b229225db094cc8e23e"},"schema_version":"1.0"},"canonical_sha256":"1a28d24937aed62aa85d2680146857f2451ca749750c7fa9aed163c9aea96338","source":{"kind":"arxiv","id":"2608.10204","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2608.10204","created_at":"2026-08-12T00:23:03Z"},{"alias_kind":"arxiv_version","alias_value":"2608.10204v1","created_at":"2026-08-12T00:23:03Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2608.10204","created_at":"2026-08-12T00:23:03Z"},{"alias_kind":"pith_short_12","alias_value":"DIUNESJXV3LC","created_at":"2026-08-12T00:23:03Z"},{"alias_kind":"pith_short_16","alias_value":"DIUNESJXV3LCVKC5","created_at":"2026-08-12T00:23:03Z"},{"alias_kind":"pith_short_8","alias_value":"DIUNESJX","created_at":"2026-08-12T00:23:03Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:DIUNESJXV3LCVKC5E2ABI2CX6J","target":"record","payload":{"canonical_record":{"source":{"id":"2608.10204","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-08-10T20:20:01Z","cross_cats_sorted":[],"title_canon_sha256":"cca53521182367838dbadcb81c688b0e7581c5448163c4ffb25bcb5fb25f631b","abstract_canon_sha256":"2167452e6d880f4d503208b54e4b8fe38341aafeefb42b229225db094cc8e23e"},"schema_version":"1.0"},"canonical_sha256":"1a28d24937aed62aa85d2680146857f2451ca749750c7fa9aed163c9aea96338","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-08-12T00:23:03.358882Z","signature_b64":"ShLuQu8XLRSu1lYsWxidM9RQwhoZeUgirZ2WXUBFbx1UwecfKT3dBNVzUXHeAVB3loPvZgoFypb1Zr66WsK+Aw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"1a28d24937aed62aa85d2680146857f2451ca749750c7fa9aed163c9aea96338","last_reissued_at":"2026-08-12T00:23:03.357210Z","signature_status":"signed_v1","first_computed_at":"2026-08-12T00:23:03.357210Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2608.10204","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":"2026-08-12T00:23:03Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"n3BkEMXKAfPDIaLB97nNiI8MYQAsMoQiNyxukJ5wE5lBpHH4XS3Gku7N9xuH57OkWH5bn/647Bj27BpXg/nmAg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-18T09:03:00.270087Z"},"content_sha256":"28a2deb5006942a916fa8ac6345d6e66be27b1df1d299f8fa70dbbac7f8d7138","schema_version":"1.0","event_id":"sha256:28a2deb5006942a916fa8ac6345d6e66be27b1df1d299f8fa70dbbac7f8d7138"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:DIUNESJXV3LCVKC5E2ABI2CX6J","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Boundary-Seeking Policy Gradient for Safe Reinforcement Learning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Chenhua Fan, Honghao Wei, Jiahui Zhu, Yuhang Zhang","submitted_at":"2026-08-10T20:20:01Z","abstract_excerpt":"Safe reinforcement learning maximizes reward subject to safety constraints. For Constrained Markov Decision Processes, the linear-programming view over occupancy measures implies that whenever the constraint is active at optimality, the optimal policy lies exactly on the constraint boundary, yet standard gradient-based methods do not exploit this structure and often settle in the feasible interior. We introduce Boundary-Seeking Policy Gradient (BSPG), a first-order method whose update combines a tangential component that improves reward while preserving cost to first order with a signed, resid"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2608.10204","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/2608.10204/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":"2026-08-12T00:23:03Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"vexOI75VgVbqpfkEm3MnvG1ajc67F+/6/7+A494f9FjpThuKmOVjBcZiHuNNWXRb97vYHw7OzK/x6YzWnJfEBQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-18T09:03:00.270892Z"},"content_sha256":"679bbc398d132e5ff5ab0fdc26920810edb2b62d15f846bdc421d7818efd6473","schema_version":"1.0","event_id":"sha256:679bbc398d132e5ff5ab0fdc26920810edb2b62d15f846bdc421d7818efd6473"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/DIUNESJXV3LCVKC5E2ABI2CX6J/bundle.json","state_url":"https://pith.science/pith/DIUNESJXV3LCVKC5E2ABI2CX6J/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/DIUNESJXV3LCVKC5E2ABI2CX6J/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":"2026-08-18T09:03:00Z","links":{"resolver":"https://pith.science/pith/DIUNESJXV3LCVKC5E2ABI2CX6J","bundle":"https://pith.science/pith/DIUNESJXV3LCVKC5E2ABI2CX6J/bundle.json","state":"https://pith.science/pith/DIUNESJXV3LCVKC5E2ABI2CX6J/state.json","well_known_bundle":"https://pith.science/.well-known/pith/DIUNESJXV3LCVKC5E2ABI2CX6J/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:DIUNESJXV3LCVKC5E2ABI2CX6J","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":"2167452e6d880f4d503208b54e4b8fe38341aafeefb42b229225db094cc8e23e","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-08-10T20:20:01Z","title_canon_sha256":"cca53521182367838dbadcb81c688b0e7581c5448163c4ffb25bcb5fb25f631b"},"schema_version":"1.0","source":{"id":"2608.10204","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2608.10204","created_at":"2026-08-12T00:23:03Z"},{"alias_kind":"arxiv_version","alias_value":"2608.10204v1","created_at":"2026-08-12T00:23:03Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2608.10204","created_at":"2026-08-12T00:23:03Z"},{"alias_kind":"pith_short_12","alias_value":"DIUNESJXV3LC","created_at":"2026-08-12T00:23:03Z"},{"alias_kind":"pith_short_16","alias_value":"DIUNESJXV3LCVKC5","created_at":"2026-08-12T00:23:03Z"},{"alias_kind":"pith_short_8","alias_value":"DIUNESJX","created_at":"2026-08-12T00:23:03Z"}],"graph_snapshots":[{"event_id":"sha256:679bbc398d132e5ff5ab0fdc26920810edb2b62d15f846bdc421d7818efd6473","target":"graph","created_at":"2026-08-12T00:23:03Z","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/2608.10204/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Safe reinforcement learning maximizes reward subject to safety constraints. For Constrained Markov Decision Processes, the linear-programming view over occupancy measures implies that whenever the constraint is active at optimality, the optimal policy lies exactly on the constraint boundary, yet standard gradient-based methods do not exploit this structure and often settle in the feasible interior. We introduce Boundary-Seeking Policy Gradient (BSPG), a first-order method whose update combines a tangential component that improves reward while preserving cost to first order with a signed, resid","authors_text":"Chenhua Fan, Honghao Wei, Jiahui Zhu, Yuhang Zhang","cross_cats":[],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-08-10T20:20:01Z","title":"Boundary-Seeking Policy Gradient for Safe Reinforcement Learning"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2608.10204","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":"sha256:28a2deb5006942a916fa8ac6345d6e66be27b1df1d299f8fa70dbbac7f8d7138","target":"record","created_at":"2026-08-12T00:23:03Z","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":"2167452e6d880f4d503208b54e4b8fe38341aafeefb42b229225db094cc8e23e","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-08-10T20:20:01Z","title_canon_sha256":"cca53521182367838dbadcb81c688b0e7581c5448163c4ffb25bcb5fb25f631b"},"schema_version":"1.0","source":{"id":"2608.10204","kind":"arxiv","version":1}},"canonical_sha256":"1a28d24937aed62aa85d2680146857f2451ca749750c7fa9aed163c9aea96338","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"1a28d24937aed62aa85d2680146857f2451ca749750c7fa9aed163c9aea96338","first_computed_at":"2026-08-12T00:23:03.357210Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-08-12T00:23:03.357210Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"ShLuQu8XLRSu1lYsWxidM9RQwhoZeUgirZ2WXUBFbx1UwecfKT3dBNVzUXHeAVB3loPvZgoFypb1Zr66WsK+Aw==","signature_status":"signed_v1","signed_at":"2026-08-12T00:23:03.358882Z","signed_message":"canonical_sha256_bytes"},"source_id":"2608.10204","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:28a2deb5006942a916fa8ac6345d6e66be27b1df1d299f8fa70dbbac7f8d7138","sha256:679bbc398d132e5ff5ab0fdc26920810edb2b62d15f846bdc421d7818efd6473"],"state_sha256":"bff62a05d6c9bd8a3e2144289b73a8e5983b5a4de3e7025809fd9f3aec0b8251"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"2GrVb3MnUe04l9i9lvsUvE0nHtyZlSV5c8HUSvyMz6aHtiIom1zLraTTq5niJTSItiNS/YvfkWiCAuirpzAsDQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-18T09:03:00.275245Z","bundle_sha256":"8c7c76c222e62c3dfcc4347660a729d018eb95ea1dafd3f882d63e17fcd5d843"}}