{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2020:DPRVM74NWXIISL4EZMY3KUMEVE","short_pith_number":"pith:DPRVM74N","schema_version":"1.0","canonical_sha256":"1be3567f8db5d0892f84cb31b55184a93ea54b3912d130b03cc7841c651b8bec","source":{"kind":"arxiv","id":"2003.12909","version":2},"attestation_state":"computed","paper":{"title":"Policy Teaching via Environment Poisoning: Training-time Adversarial Attacks against Reinforcement Learning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.CR","stat.ML"],"primary_cat":"cs.LG","authors_text":"Adish Singla, Amin Rakhsha, Goran Radanovic, Rati Devidze, Xiaojin Zhu","submitted_at":"2020-03-28T23:22:28Z","abstract_excerpt":"We study a security threat to reinforcement learning where an attacker poisons the learning environment to force the agent into executing a target policy chosen by the attacker. As a victim, we consider RL agents whose objective is to find a policy that maximizes average reward in undiscounted infinite-horizon problem settings. The attacker can manipulate the rewards or the transition dynamics in the learning environment at training-time and is interested in doing so in a stealthy manner. We propose an optimization framework for finding an \\emph{optimal stealthy attack} for different measures "},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2003.12909","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2020-03-28T23:22:28Z","cross_cats_sorted":["cs.AI","cs.CR","stat.ML"],"title_canon_sha256":"9a19f7822c6817fe6483f6e47f6cb39f579fad5262cfeb94a1d3f7cb0713fd29","abstract_canon_sha256":"e672970894382258d7fa936991599e6b3a968738a45ff413a258cb54d1f087f7"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T01:28:19.510282Z","signature_b64":"TL5Q3cqULS7qyUsks4ZCxmSJagBa5j49wW45vcxRReCX+/h0NlyNI5kcYvSdyBs2LgLcWMsb+q7/FyXXsZ8gAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"1be3567f8db5d0892f84cb31b55184a93ea54b3912d130b03cc7841c651b8bec","last_reissued_at":"2026-07-05T01:28:19.509956Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T01:28:19.509956Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Policy Teaching via Environment Poisoning: Training-time Adversarial Attacks against Reinforcement Learning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.CR","stat.ML"],"primary_cat":"cs.LG","authors_text":"Adish Singla, Amin Rakhsha, Goran Radanovic, Rati Devidze, Xiaojin Zhu","submitted_at":"2020-03-28T23:22:28Z","abstract_excerpt":"We study a security threat to reinforcement learning where an attacker poisons the learning environment to force the agent into executing a target policy chosen by the attacker. As a victim, we consider RL agents whose objective is to find a policy that maximizes average reward in undiscounted infinite-horizon problem settings. The attacker can manipulate the rewards or the transition dynamics in the learning environment at training-time and is interested in doing so in a stealthy manner. We propose an optimization framework for finding an \\emph{optimal stealthy attack} for different measures "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2003.12909","kind":"arxiv","version":2},"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/2003.12909/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"},"aliases":[{"alias_kind":"arxiv","alias_value":"2003.12909","created_at":"2026-07-05T01:28:19.510003+00:00"},{"alias_kind":"arxiv_version","alias_value":"2003.12909v2","created_at":"2026-07-05T01:28:19.510003+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2003.12909","created_at":"2026-07-05T01:28:19.510003+00:00"},{"alias_kind":"pith_short_12","alias_value":"DPRVM74NWXII","created_at":"2026-07-05T01:28:19.510003+00:00"},{"alias_kind":"pith_short_16","alias_value":"DPRVM74NWXIISL4E","created_at":"2026-07-05T01:28:19.510003+00:00"},{"alias_kind":"pith_short_8","alias_value":"DPRVM74N","created_at":"2026-07-05T01:28:19.510003+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.29012","citing_title":"The Game Changer Problem: Controlling Equilibria with Discrete Rewards","ref_index":97,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/DPRVM74NWXIISL4EZMY3KUMEVE","json":"https://pith.science/pith/DPRVM74NWXIISL4EZMY3KUMEVE.json","graph_json":"https://pith.science/api/pith-number/DPRVM74NWXIISL4EZMY3KUMEVE/graph.json","events_json":"https://pith.science/api/pith-number/DPRVM74NWXIISL4EZMY3KUMEVE/events.json","paper":"https://pith.science/paper/DPRVM74N"},"agent_actions":{"view_html":"https://pith.science/pith/DPRVM74NWXIISL4EZMY3KUMEVE","download_json":"https://pith.science/pith/DPRVM74NWXIISL4EZMY3KUMEVE.json","view_paper":"https://pith.science/paper/DPRVM74N","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2003.12909&json=true","fetch_graph":"https://pith.science/api/pith-number/DPRVM74NWXIISL4EZMY3KUMEVE/graph.json","fetch_events":"https://pith.science/api/pith-number/DPRVM74NWXIISL4EZMY3KUMEVE/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/DPRVM74NWXIISL4EZMY3KUMEVE/action/timestamp_anchor","attest_storage":"https://pith.science/pith/DPRVM74NWXIISL4EZMY3KUMEVE/action/storage_attestation","attest_author":"https://pith.science/pith/DPRVM74NWXIISL4EZMY3KUMEVE/action/author_attestation","sign_citation":"https://pith.science/pith/DPRVM74NWXIISL4EZMY3KUMEVE/action/citation_signature","submit_replication":"https://pith.science/pith/DPRVM74NWXIISL4EZMY3KUMEVE/action/replication_record"}},"created_at":"2026-07-05T01:28:19.510003+00:00","updated_at":"2026-07-05T01:28:19.510003+00:00"}