{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:DTOJ6LA4GNU76LWDO7JI5CNJKP","short_pith_number":"pith:DTOJ6LA4","schema_version":"1.0","canonical_sha256":"1cdc9f2c1c3369ff2ec377d28e89a953e4d25a72954390a2d65591bbdb4e19f9","source":{"kind":"arxiv","id":"2605.16894","version":1},"attestation_state":"computed","paper":{"title":"Beyond Safety Filtering: Control Barrier Function-Informed Reinforcement Learning for Connected and Automated Vehicles","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"Converting Control Barrier Function constraints into rewards guides multi-agent reinforcement learning to higher performance with reduced hyperparameter sensitivity in connected vehicle intersections.","cross_cats":["cs.SY","eess.SY"],"primary_cat":"cs.RO","authors_text":"Bassam Alrifaee, Jianye Xu","submitted_at":"2026-05-16T09:12:59Z","abstract_excerpt":"Reinforcement Learning (RL) uses rewards to guide learning, yet reward design is typically hand-crafted using heuristics that can be difficult to tune. We propose a Control Barrier Function (CBF)-informed reward design for Multi-Agent RL (MARL) that converts CBF constraint values under joint MARL actions into a reward signal that explicitly guides safe learning. We compare against two heuristic reward baselines in a four-way multi-lane intersection with connected and automated vehicles. Results show that our method achieves the highest task performance and is less sensitive to reward hyperpara"},"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":true,"formal_links_present":true},"canonical_record":{"source":{"id":"2605.16894","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.RO","submitted_at":"2026-05-16T09:12:59Z","cross_cats_sorted":["cs.SY","eess.SY"],"title_canon_sha256":"2a9e2e8c6745166ad0f680ab9e26e381104876c61ab53032a5784727f3070290","abstract_canon_sha256":"5f85ffaa2ee6e319efaa1018677c86535ef4427802f65f93bc42a18383be1b65"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-20T00:03:28.768734Z","signature_b64":"cYPhMCHyQU5l7hFE3cCIET+G9qGZLGQE+n94MF1fwdJtUiMt4HgrXyw/G1LuXk4YNEZE5VPCi77OmUD9Z/p6Bw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"1cdc9f2c1c3369ff2ec377d28e89a953e4d25a72954390a2d65591bbdb4e19f9","last_reissued_at":"2026-05-20T00:03:28.767960Z","signature_status":"signed_v1","first_computed_at":"2026-05-20T00:03:28.767960Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Beyond Safety Filtering: Control Barrier Function-Informed Reinforcement Learning for Connected and Automated Vehicles","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"Converting Control Barrier Function constraints into rewards guides multi-agent reinforcement learning to higher performance with reduced hyperparameter sensitivity in connected vehicle intersections.","cross_cats":["cs.SY","eess.SY"],"primary_cat":"cs.RO","authors_text":"Bassam Alrifaee, Jianye Xu","submitted_at":"2026-05-16T09:12:59Z","abstract_excerpt":"Reinforcement Learning (RL) uses rewards to guide learning, yet reward design is typically hand-crafted using heuristics that can be difficult to tune. We propose a Control Barrier Function (CBF)-informed reward design for Multi-Agent RL (MARL) that converts CBF constraint values under joint MARL actions into a reward signal that explicitly guides safe learning. We compare against two heuristic reward baselines in a four-way multi-lane intersection with connected and automated vehicles. Results show that our method achieves the highest task performance and is less sensitive to reward hyperpara"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"Results show that our method achieves the highest task performance and is less sensitive to reward hyperparameters, yielding consistently strong performance across the tested hyperparameter range.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That converting CBF constraint values under joint MARL actions into a reward signal will reliably guide safe learning without introducing new instabilities or performance trade-offs in the multi-agent intersection setting.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"CBF-informed rewards for multi-agent RL achieve higher task performance and lower sensitivity to hyperparameters than heuristic baselines in a simulated four-way intersection with connected automated vehicles.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Converting Control Barrier Function constraints into rewards guides multi-agent reinforcement learning to higher performance with reduced hyperparameter sensitivity in connected vehicle intersections.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"6ba7309c1d010f8197c85788ab2575834f4870d19e96503d77a21a5ab2a3e73b"},"source":{"id":"2605.16894","kind":"arxiv","version":1},"verdict":{"id":"21a2ade5-0107-4338-9fc5-185f5d0a3886","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-19T20:38:08.682975Z","strongest_claim":"Results show that our method achieves the highest task performance and is less sensitive to reward hyperparameters, yielding consistently strong performance across the tested hyperparameter range.","one_line_summary":"CBF-informed rewards for multi-agent RL achieve higher task performance and lower sensitivity to hyperparameters than heuristic baselines in a simulated four-way intersection with connected automated vehicles.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That converting CBF constraint values under joint MARL actions into a reward signal will reliably guide safe learning without introducing new instabilities or performance trade-offs in the multi-agent intersection setting.","pith_extraction_headline":"Converting Control Barrier Function constraints into rewards guides multi-agent reinforcement learning to higher performance with reduced hyperparameter sensitivity in connected vehicle intersections."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2605.16894/integrity.json","findings":[],"available":true,"detectors_run":[{"name":"doi_title_agreement","ran_at":"2026-05-19T21:01:19.181231Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"cited_work_retraction","ran_at":"2026-05-19T20:52:41.413749Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"doi_compliance","ran_at":"2026-05-19T20:51:10.967041Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"claim_evidence","ran_at":"2026-05-19T18:41:56.280981Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"ai_meta_artifact","ran_at":"2026-05-19T18:33:26.359628Z","status":"skipped","version":"1.0.0","findings_count":0}],"snapshot_sha256":"923fc45754ea86babb582112c625249cf9f377a01177e7b51f56e6b5c54022d4"},"references":{"count":29,"sample":[{"doi":"","year":2022,"title":"Deep reinforcement learning for autonomous driving: A survey","work_id":"fdd14588-d339-4189-97d5-6de7d0f992d3","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2023,"title":"Reward (mis) design for autonomous driving,","work_id":"a543166c-b2c2-46b6-a686-49e4ab275a99","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2019,"title":"Model-free deep reinforcement learning for urban autonomous driving,","work_id":"023e97d5-e422-4bee-9373-7ab4a5606f88","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2022,"title":"Interpretable end-to-end urban autonomous driving with latent deep reinforcement learning,","work_id":"6fdbdd49-a8a1-42be-9ed2-f5cfd5684a8f","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2017,"title":"Formulation of deep reinforcement learning architecture toward autonomous driving for on-ramp merge,","work_id":"1d8fa2b7-0ca8-4498-a50f-53873aa9e420","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":29,"snapshot_sha256":"ee3ddc91d0c70d8e68f6a7003ec8adfb94a388be1dd091f3fdb2ebd7dc671db9","internal_anchors":0},"formal_canon":{"evidence_count":2,"snapshot_sha256":"3618a5ee0c8622a988e910eb978ee37eb22349b902d2a5fbe800c699f1e3b88c"},"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":"2605.16894","created_at":"2026-05-20T00:03:28.768115+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.16894v1","created_at":"2026-05-20T00:03:28.768115+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.16894","created_at":"2026-05-20T00:03:28.768115+00:00"},{"alias_kind":"pith_short_12","alias_value":"DTOJ6LA4GNU7","created_at":"2026-05-20T00:03:28.768115+00:00"},{"alias_kind":"pith_short_16","alias_value":"DTOJ6LA4GNU76LWD","created_at":"2026-05-20T00:03:28.768115+00:00"},{"alias_kind":"pith_short_8","alias_value":"DTOJ6LA4","created_at":"2026-05-20T00:03:28.768115+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":2,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/DTOJ6LA4GNU76LWDO7JI5CNJKP","json":"https://pith.science/pith/DTOJ6LA4GNU76LWDO7JI5CNJKP.json","graph_json":"https://pith.science/api/pith-number/DTOJ6LA4GNU76LWDO7JI5CNJKP/graph.json","events_json":"https://pith.science/api/pith-number/DTOJ6LA4GNU76LWDO7JI5CNJKP/events.json","paper":"https://pith.science/paper/DTOJ6LA4"},"agent_actions":{"view_html":"https://pith.science/pith/DTOJ6LA4GNU76LWDO7JI5CNJKP","download_json":"https://pith.science/pith/DTOJ6LA4GNU76LWDO7JI5CNJKP.json","view_paper":"https://pith.science/paper/DTOJ6LA4","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.16894&json=true","fetch_graph":"https://pith.science/api/pith-number/DTOJ6LA4GNU76LWDO7JI5CNJKP/graph.json","fetch_events":"https://pith.science/api/pith-number/DTOJ6LA4GNU76LWDO7JI5CNJKP/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/DTOJ6LA4GNU76LWDO7JI5CNJKP/action/timestamp_anchor","attest_storage":"https://pith.science/pith/DTOJ6LA4GNU76LWDO7JI5CNJKP/action/storage_attestation","attest_author":"https://pith.science/pith/DTOJ6LA4GNU76LWDO7JI5CNJKP/action/author_attestation","sign_citation":"https://pith.science/pith/DTOJ6LA4GNU76LWDO7JI5CNJKP/action/citation_signature","submit_replication":"https://pith.science/pith/DTOJ6LA4GNU76LWDO7JI5CNJKP/action/replication_record"}},"created_at":"2026-05-20T00:03:28.768115+00:00","updated_at":"2026-05-20T00:03:28.768115+00:00"}