{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:DTOJ6LA4GNU76LWDO7JI5CNJKP","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"5f85ffaa2ee6e319efaa1018677c86535ef4427802f65f93bc42a18383be1b65","cross_cats_sorted":["cs.SY","eess.SY"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.RO","submitted_at":"2026-05-16T09:12:59Z","title_canon_sha256":"2a9e2e8c6745166ad0f680ab9e26e381104876c61ab53032a5784727f3070290"},"schema_version":"1.0","source":{"id":"2605.16894","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.16894","created_at":"2026-05-20T00:03:28Z"},{"alias_kind":"arxiv_version","alias_value":"2605.16894v1","created_at":"2026-05-20T00:03:28Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.16894","created_at":"2026-05-20T00:03:28Z"},{"alias_kind":"pith_short_12","alias_value":"DTOJ6LA4GNU7","created_at":"2026-05-20T00:03:28Z"},{"alias_kind":"pith_short_16","alias_value":"DTOJ6LA4GNU76LWD","created_at":"2026-05-20T00:03:28Z"},{"alias_kind":"pith_short_8","alias_value":"DTOJ6LA4","created_at":"2026-05-20T00:03:28Z"}],"graph_snapshots":[{"event_id":"sha256:a9c376ea19ee76150733eb3c10ebb613a49a1e98915e9fd38bd0ef458aebec57","target":"graph","created_at":"2026-05-20T00:03:28Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":4,"items":[{"attestation":"unclaimed","claim_id":"C1","kind":"strongest_claim","source":"verdict.strongest_claim","status":"machine_extracted","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."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","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."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","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."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"Converting Control Barrier Function constraints into rewards guides multi-agent reinforcement learning to higher performance with reduced hyperparameter sensitivity in connected vehicle intersections."}],"snapshot_sha256":"6ba7309c1d010f8197c85788ab2575834f4870d19e96503d77a21a5ab2a3e73b"},"formal_canon":{"evidence_count":2,"snapshot_sha256":"3618a5ee0c8622a988e910eb978ee37eb22349b902d2a5fbe800c699f1e3b88c"},"integrity":{"available":true,"clean":true,"detectors_run":[{"findings_count":0,"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"}],"endpoint":"/pith/2605.16894/integrity.json","findings":[],"snapshot_sha256":"923fc45754ea86babb582112c625249cf9f377a01177e7b51f56e6b5c54022d4","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"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","authors_text":"Bassam Alrifaee, Jianye Xu","cross_cats":["cs.SY","eess.SY"],"headline":"Converting Control Barrier Function constraints into rewards guides multi-agent reinforcement learning to higher performance with reduced hyperparameter sensitivity in connected vehicle intersections.","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.RO","submitted_at":"2026-05-16T09:12:59Z","title":"Beyond Safety Filtering: Control Barrier Function-Informed Reinforcement Learning for Connected and Automated Vehicles"},"references":{"count":29,"internal_anchors":0,"resolved_work":29,"sample":[{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":1,"title":"Deep reinforcement learning for autonomous driving: A survey","work_id":"fdd14588-d339-4189-97d5-6de7d0f992d3","year":2022},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":2,"title":"Reward (mis) design for autonomous driving,","work_id":"a543166c-b2c2-46b6-a686-49e4ab275a99","year":2023},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":3,"title":"Model-free deep reinforcement learning for urban autonomous driving,","work_id":"023e97d5-e422-4bee-9373-7ab4a5606f88","year":2019},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":4,"title":"Interpretable end-to-end urban autonomous driving with latent deep reinforcement learning,","work_id":"6fdbdd49-a8a1-42be-9ed2-f5cfd5684a8f","year":2022},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":5,"title":"Formulation of deep reinforcement learning architecture toward autonomous driving for on-ramp merge,","work_id":"1d8fa2b7-0ca8-4498-a50f-53873aa9e420","year":2017}],"snapshot_sha256":"ee3ddc91d0c70d8e68f6a7003ec8adfb94a388be1dd091f3fdb2ebd7dc671db9"},"source":{"id":"2605.16894","kind":"arxiv","version":1},"verdict":{"created_at":"2026-05-19T20:38:08.682975Z","id":"21a2ade5-0107-4338-9fc5-185f5d0a3886","model_set":{"reader":"grok-4.3"},"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","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.","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.","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."}},"verdict_id":"21a2ade5-0107-4338-9fc5-185f5d0a3886"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:5ba2d28e87b654aeff5ff7fa97fcacc4b3936c6c38a8ed2d9aa55a938040a5eb","target":"record","created_at":"2026-05-20T00:03:28Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"5f85ffaa2ee6e319efaa1018677c86535ef4427802f65f93bc42a18383be1b65","cross_cats_sorted":["cs.SY","eess.SY"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.RO","submitted_at":"2026-05-16T09:12:59Z","title_canon_sha256":"2a9e2e8c6745166ad0f680ab9e26e381104876c61ab53032a5784727f3070290"},"schema_version":"1.0","source":{"id":"2605.16894","kind":"arxiv","version":1}},"canonical_sha256":"1cdc9f2c1c3369ff2ec377d28e89a953e4d25a72954390a2d65591bbdb4e19f9","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"1cdc9f2c1c3369ff2ec377d28e89a953e4d25a72954390a2d65591bbdb4e19f9","first_computed_at":"2026-05-20T00:03:28.767960Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-20T00:03:28.767960Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"cYPhMCHyQU5l7hFE3cCIET+G9qGZLGQE+n94MF1fwdJtUiMt4HgrXyw/G1LuXk4YNEZE5VPCi77OmUD9Z/p6Bw==","signature_status":"signed_v1","signed_at":"2026-05-20T00:03:28.768734Z","signed_message":"canonical_sha256_bytes"},"source_id":"2605.16894","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:5ba2d28e87b654aeff5ff7fa97fcacc4b3936c6c38a8ed2d9aa55a938040a5eb","sha256:a9c376ea19ee76150733eb3c10ebb613a49a1e98915e9fd38bd0ef458aebec57"],"state_sha256":"1249def6695e0663a0c494b57c47f988150218028ef4f3ca33209d19c80f441b"}