{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2020:EFAVD6SNKGEOAV3H5UDB3U67DU","short_pith_number":"pith:EFAVD6SN","schema_version":"1.0","canonical_sha256":"214151fa4d5188e05767ed061dd3df1d1469e13abc5373ef8b188ae392a38c95","source":{"kind":"arxiv","id":"2009.12612","version":2},"attestation_state":"computed","paper":{"title":"Neurosymbolic Reinforcement Learning with Formally Verified Exploration","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["stat.ML"],"primary_cat":"cs.LG","authors_text":"Abhinav Verma, Greg Anderson, Isil Dillig, Swarat Chaudhuri","submitted_at":"2020-09-26T14:51:04Z","abstract_excerpt":"We present Revel, a partially neural reinforcement learning (RL) framework for provably safe exploration in continuous state and action spaces. A key challenge for provably safe deep RL is that repeatedly verifying neural networks within a learning loop is computationally infeasible. We address this challenge using two policy classes: a general, neurosymbolic class with approximate gradients and a more restricted class of symbolic policies that allows efficient verification. Our learning algorithm is a mirror descent over policies: in each iteration, it safely lifts a symbolic policy into the "},"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":"2009.12612","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2020-09-26T14:51:04Z","cross_cats_sorted":["stat.ML"],"title_canon_sha256":"3e6b2276e8a5e20f0a45cc7c06879ac3ee96a8119b63fa08ce48e839860ddfe8","abstract_canon_sha256":"2c2639380e0ba72a90624d2b8d7b18665131d1ffc031955f31ddb2f06b37a22a"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T01:46:03.003540Z","signature_b64":"0UgADEe9rBCmR0xyGQe77PPl9tlVEjoS5mozSCCRfZYWDa4kYGBQycP8OEhmdh4FTQVZrP6vAzzFqiU8Qg2BCA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"214151fa4d5188e05767ed061dd3df1d1469e13abc5373ef8b188ae392a38c95","last_reissued_at":"2026-07-05T01:46:03.002972Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T01:46:03.002972Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Neurosymbolic Reinforcement Learning with Formally Verified Exploration","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["stat.ML"],"primary_cat":"cs.LG","authors_text":"Abhinav Verma, Greg Anderson, Isil Dillig, Swarat Chaudhuri","submitted_at":"2020-09-26T14:51:04Z","abstract_excerpt":"We present Revel, a partially neural reinforcement learning (RL) framework for provably safe exploration in continuous state and action spaces. A key challenge for provably safe deep RL is that repeatedly verifying neural networks within a learning loop is computationally infeasible. We address this challenge using two policy classes: a general, neurosymbolic class with approximate gradients and a more restricted class of symbolic policies that allows efficient verification. Our learning algorithm is a mirror descent over policies: in each iteration, it safely lifts a symbolic policy into the "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2009.12612","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/2009.12612/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":"2009.12612","created_at":"2026-07-05T01:46:03.003032+00:00"},{"alias_kind":"arxiv_version","alias_value":"2009.12612v2","created_at":"2026-07-05T01:46:03.003032+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2009.12612","created_at":"2026-07-05T01:46:03.003032+00:00"},{"alias_kind":"pith_short_12","alias_value":"EFAVD6SNKGEO","created_at":"2026-07-05T01:46:03.003032+00:00"},{"alias_kind":"pith_short_16","alias_value":"EFAVD6SNKGEOAV3H","created_at":"2026-07-05T01:46:03.003032+00:00"},{"alias_kind":"pith_short_8","alias_value":"EFAVD6SN","created_at":"2026-07-05T01:46:03.003032+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2502.09233","citing_title":"Commonsense Reasoning-Aided Autonomous Vehicle Systems","ref_index":1,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/EFAVD6SNKGEOAV3H5UDB3U67DU","json":"https://pith.science/pith/EFAVD6SNKGEOAV3H5UDB3U67DU.json","graph_json":"https://pith.science/api/pith-number/EFAVD6SNKGEOAV3H5UDB3U67DU/graph.json","events_json":"https://pith.science/api/pith-number/EFAVD6SNKGEOAV3H5UDB3U67DU/events.json","paper":"https://pith.science/paper/EFAVD6SN"},"agent_actions":{"view_html":"https://pith.science/pith/EFAVD6SNKGEOAV3H5UDB3U67DU","download_json":"https://pith.science/pith/EFAVD6SNKGEOAV3H5UDB3U67DU.json","view_paper":"https://pith.science/paper/EFAVD6SN","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2009.12612&json=true","fetch_graph":"https://pith.science/api/pith-number/EFAVD6SNKGEOAV3H5UDB3U67DU/graph.json","fetch_events":"https://pith.science/api/pith-number/EFAVD6SNKGEOAV3H5UDB3U67DU/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/EFAVD6SNKGEOAV3H5UDB3U67DU/action/timestamp_anchor","attest_storage":"https://pith.science/pith/EFAVD6SNKGEOAV3H5UDB3U67DU/action/storage_attestation","attest_author":"https://pith.science/pith/EFAVD6SNKGEOAV3H5UDB3U67DU/action/author_attestation","sign_citation":"https://pith.science/pith/EFAVD6SNKGEOAV3H5UDB3U67DU/action/citation_signature","submit_replication":"https://pith.science/pith/EFAVD6SNKGEOAV3H5UDB3U67DU/action/replication_record"}},"created_at":"2026-07-05T01:46:03.003032+00:00","updated_at":"2026-07-05T01:46:03.003032+00:00"}