{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2018:YOUXTHHEJ72NQ7S4FLQ7EGA33Z","short_pith_number":"pith:YOUXTHHE","schema_version":"1.0","canonical_sha256":"c3a9799ce44ff4d87e5c2ae1f2181bde6770a205d6e2150766e9d859c880eb61","source":{"kind":"arxiv","id":"1805.08313","version":2},"attestation_state":"computed","paper":{"title":"Learning Safe Policies with Expert Guidance","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","stat.ML"],"primary_cat":"cs.LG","authors_text":"Doina Precup, Fa Wu, Jessie Huang, Yang Cai","submitted_at":"2018-05-21T22:40:07Z","abstract_excerpt":"We propose a framework for ensuring safe behavior of a reinforcement learning agent when the reward function may be difficult to specify. In order to do this, we rely on the existence of demonstrations from expert policies, and we provide a theoretical framework for the agent to optimize in the space of rewards consistent with its existing knowledge. We propose two methods to solve the resulting optimization: an exact ellipsoid-based method and a method in the spirit of the \"follow-the-perturbed-leader\" algorithm. Our experiments demonstrate the behavior of our algorithm in both discrete and c"},"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":"1805.08313","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2018-05-21T22:40:07Z","cross_cats_sorted":["cs.AI","stat.ML"],"title_canon_sha256":"2e7755388a1c3fbdc166926bb789eeca5ac4dbe03b5b770bfd9fbe9e9d7a6d1d","abstract_canon_sha256":"1875a140d1ebf49f3c3ee82cd15d02449aff3f750e4fa078cbcdbfb0d71a418a"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-18T00:00:13.169104Z","signature_b64":"dKs1/McCf3Rw6VhnHAUxHhdbNkTiUk7grTxXGljWUQgmmfe6OG0bNc79EG5eQiQiw9UvCG8G1K261Noq14tHCg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"c3a9799ce44ff4d87e5c2ae1f2181bde6770a205d6e2150766e9d859c880eb61","last_reissued_at":"2026-05-18T00:00:13.168474Z","signature_status":"signed_v1","first_computed_at":"2026-05-18T00:00:13.168474Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Learning Safe Policies with Expert Guidance","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","stat.ML"],"primary_cat":"cs.LG","authors_text":"Doina Precup, Fa Wu, Jessie Huang, Yang Cai","submitted_at":"2018-05-21T22:40:07Z","abstract_excerpt":"We propose a framework for ensuring safe behavior of a reinforcement learning agent when the reward function may be difficult to specify. In order to do this, we rely on the existence of demonstrations from expert policies, and we provide a theoretical framework for the agent to optimize in the space of rewards consistent with its existing knowledge. We propose two methods to solve the resulting optimization: an exact ellipsoid-based method and a method in the spirit of the \"follow-the-perturbed-leader\" algorithm. Our experiments demonstrate the behavior of our algorithm in both discrete and c"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1805.08313","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":""},"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":"1805.08313","created_at":"2026-05-18T00:00:13.168563+00:00"},{"alias_kind":"arxiv_version","alias_value":"1805.08313v2","created_at":"2026-05-18T00:00:13.168563+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1805.08313","created_at":"2026-05-18T00:00:13.168563+00:00"},{"alias_kind":"pith_short_12","alias_value":"YOUXTHHEJ72N","created_at":"2026-05-18T12:33:04.347982+00:00"},{"alias_kind":"pith_short_16","alias_value":"YOUXTHHEJ72NQ7S4","created_at":"2026-05-18T12:33:04.347982+00:00"},{"alias_kind":"pith_short_8","alias_value":"YOUXTHHE","created_at":"2026-05-18T12:33:04.347982+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"1906.10099","citing_title":"DynoPlan: Combining Motion Planning and Deep Neural Network based Controllers for Safe HRL","ref_index":12,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/YOUXTHHEJ72NQ7S4FLQ7EGA33Z","json":"https://pith.science/pith/YOUXTHHEJ72NQ7S4FLQ7EGA33Z.json","graph_json":"https://pith.science/api/pith-number/YOUXTHHEJ72NQ7S4FLQ7EGA33Z/graph.json","events_json":"https://pith.science/api/pith-number/YOUXTHHEJ72NQ7S4FLQ7EGA33Z/events.json","paper":"https://pith.science/paper/YOUXTHHE"},"agent_actions":{"view_html":"https://pith.science/pith/YOUXTHHEJ72NQ7S4FLQ7EGA33Z","download_json":"https://pith.science/pith/YOUXTHHEJ72NQ7S4FLQ7EGA33Z.json","view_paper":"https://pith.science/paper/YOUXTHHE","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=1805.08313&json=true","fetch_graph":"https://pith.science/api/pith-number/YOUXTHHEJ72NQ7S4FLQ7EGA33Z/graph.json","fetch_events":"https://pith.science/api/pith-number/YOUXTHHEJ72NQ7S4FLQ7EGA33Z/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/YOUXTHHEJ72NQ7S4FLQ7EGA33Z/action/timestamp_anchor","attest_storage":"https://pith.science/pith/YOUXTHHEJ72NQ7S4FLQ7EGA33Z/action/storage_attestation","attest_author":"https://pith.science/pith/YOUXTHHEJ72NQ7S4FLQ7EGA33Z/action/author_attestation","sign_citation":"https://pith.science/pith/YOUXTHHEJ72NQ7S4FLQ7EGA33Z/action/citation_signature","submit_replication":"https://pith.science/pith/YOUXTHHEJ72NQ7S4FLQ7EGA33Z/action/replication_record"}},"created_at":"2026-05-18T00:00:13.168563+00:00","updated_at":"2026-05-18T00:00:13.168563+00:00"}