{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:SAD3A72KPMTIA4G4QHLU6KABHR","short_pith_number":"pith:SAD3A72K","schema_version":"1.0","canonical_sha256":"9007b07f4a7b268070dc81d74f28013c4550a19fd7136fd269a020d898b0703f","source":{"kind":"arxiv","id":"2605.19771","version":1},"attestation_state":"computed","paper":{"title":"Beyond Imitation: Learning Safe End-to-End Autonomous Driving from Hard Negatives","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":["cs.CV"],"primary_cat":"cs.RO","authors_text":"Guang Chen, Hangjun Ye, Haochen Tian, Junli Wang, Kun Ma, Long Chen, Qichao Zhang, Xueyi Liu, Zebin Xing, Zhihua Hua","submitted_at":"2026-05-19T12:41:47Z","abstract_excerpt":"Existing imitation learning methods for end-to-end autonomous driving predominantly learn from successful demonstrations by minimizing geometric deviations from expert trajectories. This paradigm implicitly assumes that spatial proximity implies behavioral safety, leading to a critical objective mismatch: trajectories with nearly identical imitation losses may exhibit drastically different safety outcomes, where one remains recoverable while the other results in collision. To address this limitation, we propose BeyondDrive, a failure-aware imitation learning framework that jointly learns from "},"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":"2605.19771","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.RO","submitted_at":"2026-05-19T12:41:47Z","cross_cats_sorted":["cs.CV"],"title_canon_sha256":"37c2657b68e55fab94d5a371803d70ba0c718528b258895fad104d9c9e2c4b5b","abstract_canon_sha256":"65d4add9c8179c3bf7d1e9584eabf1606f94a0e75dfe8677d3a473bea9c0bac8"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-20T01:06:13.028554Z","signature_b64":"AM89b0vWEnDQFgNV5o4z6zTiL/pNroQU7rwATBalUqmeD184JTTIwcwCIkPxRxhruxDM+zOT37jDVLqh68oPBA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"9007b07f4a7b268070dc81d74f28013c4550a19fd7136fd269a020d898b0703f","last_reissued_at":"2026-05-20T01:06:13.027859Z","signature_status":"signed_v1","first_computed_at":"2026-05-20T01:06:13.027859Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Beyond Imitation: Learning Safe End-to-End Autonomous Driving from Hard Negatives","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":["cs.CV"],"primary_cat":"cs.RO","authors_text":"Guang Chen, Hangjun Ye, Haochen Tian, Junli Wang, Kun Ma, Long Chen, Qichao Zhang, Xueyi Liu, Zebin Xing, Zhihua Hua","submitted_at":"2026-05-19T12:41:47Z","abstract_excerpt":"Existing imitation learning methods for end-to-end autonomous driving predominantly learn from successful demonstrations by minimizing geometric deviations from expert trajectories. This paradigm implicitly assumes that spatial proximity implies behavioral safety, leading to a critical objective mismatch: trajectories with nearly identical imitation losses may exhibit drastically different safety outcomes, where one remains recoverable while the other results in collision. To address this limitation, we propose BeyondDrive, a failure-aware imitation learning framework that jointly learns from "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.19771","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/2605.19771/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":"2605.19771","created_at":"2026-05-20T01:06:13.027990+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.19771v1","created_at":"2026-05-20T01:06:13.027990+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.19771","created_at":"2026-05-20T01:06:13.027990+00:00"},{"alias_kind":"pith_short_12","alias_value":"SAD3A72KPMTI","created_at":"2026-05-20T01:06:13.027990+00:00"},{"alias_kind":"pith_short_16","alias_value":"SAD3A72KPMTIA4G4","created_at":"2026-05-20T01:06:13.027990+00:00"},{"alias_kind":"pith_short_8","alias_value":"SAD3A72K","created_at":"2026-05-20T01:06:13.027990+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/SAD3A72KPMTIA4G4QHLU6KABHR","json":"https://pith.science/pith/SAD3A72KPMTIA4G4QHLU6KABHR.json","graph_json":"https://pith.science/api/pith-number/SAD3A72KPMTIA4G4QHLU6KABHR/graph.json","events_json":"https://pith.science/api/pith-number/SAD3A72KPMTIA4G4QHLU6KABHR/events.json","paper":"https://pith.science/paper/SAD3A72K"},"agent_actions":{"view_html":"https://pith.science/pith/SAD3A72KPMTIA4G4QHLU6KABHR","download_json":"https://pith.science/pith/SAD3A72KPMTIA4G4QHLU6KABHR.json","view_paper":"https://pith.science/paper/SAD3A72K","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.19771&json=true","fetch_graph":"https://pith.science/api/pith-number/SAD3A72KPMTIA4G4QHLU6KABHR/graph.json","fetch_events":"https://pith.science/api/pith-number/SAD3A72KPMTIA4G4QHLU6KABHR/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/SAD3A72KPMTIA4G4QHLU6KABHR/action/timestamp_anchor","attest_storage":"https://pith.science/pith/SAD3A72KPMTIA4G4QHLU6KABHR/action/storage_attestation","attest_author":"https://pith.science/pith/SAD3A72KPMTIA4G4QHLU6KABHR/action/author_attestation","sign_citation":"https://pith.science/pith/SAD3A72KPMTIA4G4QHLU6KABHR/action/citation_signature","submit_replication":"https://pith.science/pith/SAD3A72KPMTIA4G4QHLU6KABHR/action/replication_record"}},"created_at":"2026-05-20T01:06:13.027990+00:00","updated_at":"2026-05-20T01:06:13.027990+00:00"}