{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:QXJSVAQQC5LRVN375F4FO4NU2Q","short_pith_number":"pith:QXJSVAQQ","schema_version":"1.0","canonical_sha256":"85d32a821017571ab77fe9785771b4d42ec5dcc61adee3d47585ca677366c5db","source":{"kind":"arxiv","id":"2606.00270","version":1},"attestation_state":"computed","paper":{"title":"Robust Shielding for Safe Reinforcement Learning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG","cs.LO"],"primary_cat":"cs.AI","authors_text":"Alessandro Abate, Edwin Hamel-De le Court, Francesco Belardinelli, Francesco Fabiano, Thom Badings","submitted_at":"2026-05-29T19:01:12Z","abstract_excerpt":"Shielding is an effective approach to formally guarantee the safety of reinforcement learning agents in Markov decision processes (MDPs). However, existing shielding techniques typically assume knowledge of the safety-relevant transition dynamics - a requirement that is seldom met in practice. To address this limitation, we introduce a novel shielding framework for robust MDPs (RMDPs), i.e., MDPs with sets of transition probabilities. We define safety as the satisfaction of a linear temporal logic (LTL) formula with a certain threshold probability under the worst-case transition probabilities "},"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":"2606.00270","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2026-05-29T19:01:12Z","cross_cats_sorted":["cs.LG","cs.LO"],"title_canon_sha256":"6e6f7d90062120e55b5c92c17a7a76d0fc1527f065bf721bac275917a9f2849b","abstract_canon_sha256":"9c1cf4057598820f95eef45ac4b920e27422bbf4dddd5208f14b96a796e3f772"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-02T01:03:23.381289Z","signature_b64":"R0ujqa3Pm//2Zbp6ADu4rUIgv0JnAizuUudOHDnXWuiXkK0fabFAPg0U/80elkBNsQoMob4v6so5QxkkySj2DQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"85d32a821017571ab77fe9785771b4d42ec5dcc61adee3d47585ca677366c5db","last_reissued_at":"2026-06-02T01:03:23.380911Z","signature_status":"signed_v1","first_computed_at":"2026-06-02T01:03:23.380911Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Robust Shielding for Safe Reinforcement Learning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG","cs.LO"],"primary_cat":"cs.AI","authors_text":"Alessandro Abate, Edwin Hamel-De le Court, Francesco Belardinelli, Francesco Fabiano, Thom Badings","submitted_at":"2026-05-29T19:01:12Z","abstract_excerpt":"Shielding is an effective approach to formally guarantee the safety of reinforcement learning agents in Markov decision processes (MDPs). However, existing shielding techniques typically assume knowledge of the safety-relevant transition dynamics - a requirement that is seldom met in practice. To address this limitation, we introduce a novel shielding framework for robust MDPs (RMDPs), i.e., MDPs with sets of transition probabilities. We define safety as the satisfaction of a linear temporal logic (LTL) formula with a certain threshold probability under the worst-case transition probabilities "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.00270","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/2606.00270/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":"2606.00270","created_at":"2026-06-02T01:03:23.380967+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.00270v1","created_at":"2026-06-02T01:03:23.380967+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.00270","created_at":"2026-06-02T01:03:23.380967+00:00"},{"alias_kind":"pith_short_12","alias_value":"QXJSVAQQC5LR","created_at":"2026-06-02T01:03:23.380967+00:00"},{"alias_kind":"pith_short_16","alias_value":"QXJSVAQQC5LRVN37","created_at":"2026-06-02T01:03:23.380967+00:00"},{"alias_kind":"pith_short_8","alias_value":"QXJSVAQQ","created_at":"2026-06-02T01:03:23.380967+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/QXJSVAQQC5LRVN375F4FO4NU2Q","json":"https://pith.science/pith/QXJSVAQQC5LRVN375F4FO4NU2Q.json","graph_json":"https://pith.science/api/pith-number/QXJSVAQQC5LRVN375F4FO4NU2Q/graph.json","events_json":"https://pith.science/api/pith-number/QXJSVAQQC5LRVN375F4FO4NU2Q/events.json","paper":"https://pith.science/paper/QXJSVAQQ"},"agent_actions":{"view_html":"https://pith.science/pith/QXJSVAQQC5LRVN375F4FO4NU2Q","download_json":"https://pith.science/pith/QXJSVAQQC5LRVN375F4FO4NU2Q.json","view_paper":"https://pith.science/paper/QXJSVAQQ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.00270&json=true","fetch_graph":"https://pith.science/api/pith-number/QXJSVAQQC5LRVN375F4FO4NU2Q/graph.json","fetch_events":"https://pith.science/api/pith-number/QXJSVAQQC5LRVN375F4FO4NU2Q/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/QXJSVAQQC5LRVN375F4FO4NU2Q/action/timestamp_anchor","attest_storage":"https://pith.science/pith/QXJSVAQQC5LRVN375F4FO4NU2Q/action/storage_attestation","attest_author":"https://pith.science/pith/QXJSVAQQC5LRVN375F4FO4NU2Q/action/author_attestation","sign_citation":"https://pith.science/pith/QXJSVAQQC5LRVN375F4FO4NU2Q/action/citation_signature","submit_replication":"https://pith.science/pith/QXJSVAQQC5LRVN375F4FO4NU2Q/action/replication_record"}},"created_at":"2026-06-02T01:03:23.380967+00:00","updated_at":"2026-06-02T01:03:23.380967+00:00"}