{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2022:HBIADRTMMDX3TGZNFFAEVKRZO2","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":"27e5412378a79b95512ae7b11f75f742a17808d441ccd5347018a352e56afea8","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.LG","submitted_at":"2022-11-18T23:12:24Z","title_canon_sha256":"db20b888e23243378170c9cc42a76bf7423c3b58746c549dd62124d301468b43"},"schema_version":"1.0","source":{"id":"2211.10530","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2211.10530","created_at":"2026-07-05T05:17:31Z"},{"alias_kind":"arxiv_version","alias_value":"2211.10530v1","created_at":"2026-07-05T05:17:31Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2211.10530","created_at":"2026-07-05T05:17:31Z"},{"alias_kind":"pith_short_12","alias_value":"HBIADRTMMDX3","created_at":"2026-07-05T05:17:31Z"},{"alias_kind":"pith_short_16","alias_value":"HBIADRTMMDX3TGZN","created_at":"2026-07-05T05:17:31Z"},{"alias_kind":"pith_short_8","alias_value":"HBIADRTM","created_at":"2026-07-05T05:17:31Z"}],"graph_snapshots":[{"event_id":"sha256:a4338f2f5cd28b6f8cc7e4eb9271fa0f21a00a4cdb088fc98840a81d22f4edf9","target":"graph","created_at":"2026-07-05T05:17:31Z","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/2211.10530/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"We propose a provable defense mechanism against backdoor policies in reinforcement learning under subspace trigger assumption. A backdoor policy is a security threat where an adversary publishes a seemingly well-behaved policy which in fact allows hidden triggers. During deployment, the adversary can modify observed states in a particular way to trigger unexpected actions and harm the agent. We assume the agent does not have the resources to re-train a good policy. Instead, our defense mechanism sanitizes the backdoor policy by projecting observed states to a 'safe subspace', estimated from a ","authors_text":"Adish Singla, Shubham Kumar Bharti, Xiaojin Zhu, Xuezhou Zhang","cross_cats":["cs.AI"],"headline":"","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.LG","submitted_at":"2022-11-18T23:12:24Z","title":"Provable Defense against Backdoor Policies in Reinforcement Learning"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2211.10530","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":"sha256:5080175c1c199ebc436609aca6927568dde4016a279b1b146cd75dfdf69dd1d1","target":"record","created_at":"2026-07-05T05:17:31Z","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":"27e5412378a79b95512ae7b11f75f742a17808d441ccd5347018a352e56afea8","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.LG","submitted_at":"2022-11-18T23:12:24Z","title_canon_sha256":"db20b888e23243378170c9cc42a76bf7423c3b58746c549dd62124d301468b43"},"schema_version":"1.0","source":{"id":"2211.10530","kind":"arxiv","version":1}},"canonical_sha256":"385001c66c60efb99b2d29404aaa3976bb0de94151cbf92108a1c63ad6668811","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"385001c66c60efb99b2d29404aaa3976bb0de94151cbf92108a1c63ad6668811","first_computed_at":"2026-07-05T05:17:31.436520Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T05:17:31.436520Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"8tFsdpMN/BqbnKNEmsvBuTRnEZeHpFz7T205FVNrQK0LK63vOfO03lproKvjr8U0P6pMkaED0HgU9yEqELyBBA==","signature_status":"signed_v1","signed_at":"2026-07-05T05:17:31.436898Z","signed_message":"canonical_sha256_bytes"},"source_id":"2211.10530","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:5080175c1c199ebc436609aca6927568dde4016a279b1b146cd75dfdf69dd1d1","sha256:a4338f2f5cd28b6f8cc7e4eb9271fa0f21a00a4cdb088fc98840a81d22f4edf9"],"state_sha256":"317cb3fd91a1ab88abf92022393c272d5e6278506611678c49be8b414b2e6931"}