{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:CGS4OZ6HT6K62XUWUJ2GJ3RUZD","short_pith_number":"pith:CGS4OZ6H","canonical_record":{"source":{"id":"2405.05439","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.RO","submitted_at":"2024-05-08T22:00:35Z","cross_cats_sorted":["cs.AI","cs.LG","stat.AP"],"title_canon_sha256":"4b00b427ab352348ff5dc2688751f768aaa022c0d1e9983bec5abf7c71dae084","abstract_canon_sha256":"31f5a15107f6bdef418f1709036e675dc812bd22c106aeae652c6ef87f542fde"},"schema_version":"1.0"},"canonical_sha256":"11a5c767c79f95ed5e96a27464ee34c8ddc7ff7d07bf689ec5fb3f851b15a07f","source":{"kind":"arxiv","id":"2405.05439","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2405.05439","created_at":"2026-07-05T09:24:34Z"},{"alias_kind":"arxiv_version","alias_value":"2405.05439v2","created_at":"2026-07-05T09:24:34Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2405.05439","created_at":"2026-07-05T09:24:34Z"},{"alias_kind":"pith_short_12","alias_value":"CGS4OZ6HT6K6","created_at":"2026-07-05T09:24:34Z"},{"alias_kind":"pith_short_16","alias_value":"CGS4OZ6HT6K62XUW","created_at":"2026-07-05T09:24:34Z"},{"alias_kind":"pith_short_8","alias_value":"CGS4OZ6H","created_at":"2026-07-05T09:24:34Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:CGS4OZ6HT6K62XUWUJ2GJ3RUZD","target":"record","payload":{"canonical_record":{"source":{"id":"2405.05439","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.RO","submitted_at":"2024-05-08T22:00:35Z","cross_cats_sorted":["cs.AI","cs.LG","stat.AP"],"title_canon_sha256":"4b00b427ab352348ff5dc2688751f768aaa022c0d1e9983bec5abf7c71dae084","abstract_canon_sha256":"31f5a15107f6bdef418f1709036e675dc812bd22c106aeae652c6ef87f542fde"},"schema_version":"1.0"},"canonical_sha256":"11a5c767c79f95ed5e96a27464ee34c8ddc7ff7d07bf689ec5fb3f851b15a07f","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:24:34.664142Z","signature_b64":"rT/275DjRBPj/ifxLFFgQg/xXlAbCkNdD/ysWUScHWU0kR2XTpkjWw052zrKpbJbuTy7rup4h6bdn2Bf0y6YDg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"11a5c767c79f95ed5e96a27464ee34c8ddc7ff7d07bf689ec5fb3f851b15a07f","last_reissued_at":"2026-07-05T09:24:34.663646Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:24:34.663646Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2405.05439","source_version":2,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T09:24:34Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"20WPWyqsUuH0x+F6RP/RgWrgswSzRghB9Tk9jJ1sbh99E6I1L9tZyUC69K3mHTCnPfbHPO9j9oyz7qSd9empCQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-03T19:43:41.459140Z"},"content_sha256":"bfd6208fcb65c6655aac0614e2d7a796cb7ee96da5b0275740f2a27f349d3100","schema_version":"1.0","event_id":"sha256:bfd6208fcb65c6655aac0614e2d7a796cb7ee96da5b0275740f2a27f349d3100"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:CGS4OZ6HT6K62XUWUJ2GJ3RUZD","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"How Generalizable Is My Behavior Cloning Policy? A Statistical Approach to Trustworthy Performance Evaluation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.LG","stat.AP"],"primary_cat":"cs.RO","authors_text":"Haruki Nishimura, Joseph A. Vincent, Mac Schwager, Masha Itkina, Paarth Shah, Thomas Kollar","submitted_at":"2024-05-08T22:00:35Z","abstract_excerpt":"With the rise of stochastic generative models in robot policy learning, end-to-end visuomotor policies are increasingly successful at solving complex tasks by learning from human demonstrations. Nevertheless, since real-world evaluation costs afford users only a small number of policy rollouts, it remains a challenge to accurately gauge the performance of such policies. This is exacerbated by distribution shifts causing unpredictable changes in performance during deployment. To rigorously evaluate behavior cloning policies, we present a framework that provides a tight lower-bound on robot perf"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2405.05439","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/2405.05439/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"},"verdict_id":null},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T09:24:34Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"/XSvj1pyfO6D0CkNUBjZW2IXFyW2F+NO4RQT/KMU1tMojkgDPT/WUKNW6Ry554DGgtoz/CSlfzCSL0coZ6uMAA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-03T19:43:41.459688Z"},"content_sha256":"963e00736882047f48e2e6564147342f09e5f3a4c28e5e7b559df26f684a0d0d","schema_version":"1.0","event_id":"sha256:963e00736882047f48e2e6564147342f09e5f3a4c28e5e7b559df26f684a0d0d"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/CGS4OZ6HT6K62XUWUJ2GJ3RUZD/bundle.json","state_url":"https://pith.science/pith/CGS4OZ6HT6K62XUWUJ2GJ3RUZD/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/CGS4OZ6HT6K62XUWUJ2GJ3RUZD/bundle.json","status":"primary"}],"public_keys":[{"key_id":"pith-v1-2026-05","algorithm":"ed25519","format":"raw","public_key_b64":"stVStoiQhXFxp4s2pdzPNoqVNBMojDU/fJ2db5S3CbM=","public_key_hex":"b2d552b68890857171a78b36a5dccf368a953413288c353f7c9d9d6f94b709b3","fingerprint_sha256_b32_first128bits":"RVFV5Z2OI2J3ZUO7ERDEBCYNKS","fingerprint_sha256_hex":"8d4b5ee74e4693bcd1df2446408b0d54","rotates_at":null,"url":"https://pith.science/pith-signing-key.json","notes":"Pith uses this Ed25519 key to sign canonical record SHA-256 digests. Verify with: ed25519_verify(public_key, message=canonical_sha256_bytes, signature=base64decode(signature_b64))."}],"merge_version":"pith-open-graph-merge-v1","built_at":"2026-08-03T19:43:41Z","links":{"resolver":"https://pith.science/pith/CGS4OZ6HT6K62XUWUJ2GJ3RUZD","bundle":"https://pith.science/pith/CGS4OZ6HT6K62XUWUJ2GJ3RUZD/bundle.json","state":"https://pith.science/pith/CGS4OZ6HT6K62XUWUJ2GJ3RUZD/state.json","well_known_bundle":"https://pith.science/.well-known/pith/CGS4OZ6HT6K62XUWUJ2GJ3RUZD/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:CGS4OZ6HT6K62XUWUJ2GJ3RUZD","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":"31f5a15107f6bdef418f1709036e675dc812bd22c106aeae652c6ef87f542fde","cross_cats_sorted":["cs.AI","cs.LG","stat.AP"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.RO","submitted_at":"2024-05-08T22:00:35Z","title_canon_sha256":"4b00b427ab352348ff5dc2688751f768aaa022c0d1e9983bec5abf7c71dae084"},"schema_version":"1.0","source":{"id":"2405.05439","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2405.05439","created_at":"2026-07-05T09:24:34Z"},{"alias_kind":"arxiv_version","alias_value":"2405.05439v2","created_at":"2026-07-05T09:24:34Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2405.05439","created_at":"2026-07-05T09:24:34Z"},{"alias_kind":"pith_short_12","alias_value":"CGS4OZ6HT6K6","created_at":"2026-07-05T09:24:34Z"},{"alias_kind":"pith_short_16","alias_value":"CGS4OZ6HT6K62XUW","created_at":"2026-07-05T09:24:34Z"},{"alias_kind":"pith_short_8","alias_value":"CGS4OZ6H","created_at":"2026-07-05T09:24:34Z"}],"graph_snapshots":[{"event_id":"sha256:963e00736882047f48e2e6564147342f09e5f3a4c28e5e7b559df26f684a0d0d","target":"graph","created_at":"2026-07-05T09:24:34Z","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":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2405.05439/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"With the rise of stochastic generative models in robot policy learning, end-to-end visuomotor policies are increasingly successful at solving complex tasks by learning from human demonstrations. Nevertheless, since real-world evaluation costs afford users only a small number of policy rollouts, it remains a challenge to accurately gauge the performance of such policies. This is exacerbated by distribution shifts causing unpredictable changes in performance during deployment. To rigorously evaluate behavior cloning policies, we present a framework that provides a tight lower-bound on robot perf","authors_text":"Haruki Nishimura, Joseph A. Vincent, Mac Schwager, Masha Itkina, Paarth Shah, Thomas Kollar","cross_cats":["cs.AI","cs.LG","stat.AP"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.RO","submitted_at":"2024-05-08T22:00:35Z","title":"How Generalizable Is My Behavior Cloning Policy? A Statistical Approach to Trustworthy Performance Evaluation"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2405.05439","kind":"arxiv","version":2},"verdict":{"created_at":null,"id":null,"model_set":{},"one_line_summary":"","pipeline_version":null,"pith_extraction_headline":"","strongest_claim":"","weakest_assumption":""}},"verdict_id":null}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:bfd6208fcb65c6655aac0614e2d7a796cb7ee96da5b0275740f2a27f349d3100","target":"record","created_at":"2026-07-05T09:24:34Z","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":"31f5a15107f6bdef418f1709036e675dc812bd22c106aeae652c6ef87f542fde","cross_cats_sorted":["cs.AI","cs.LG","stat.AP"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.RO","submitted_at":"2024-05-08T22:00:35Z","title_canon_sha256":"4b00b427ab352348ff5dc2688751f768aaa022c0d1e9983bec5abf7c71dae084"},"schema_version":"1.0","source":{"id":"2405.05439","kind":"arxiv","version":2}},"canonical_sha256":"11a5c767c79f95ed5e96a27464ee34c8ddc7ff7d07bf689ec5fb3f851b15a07f","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"11a5c767c79f95ed5e96a27464ee34c8ddc7ff7d07bf689ec5fb3f851b15a07f","first_computed_at":"2026-07-05T09:24:34.663646Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T09:24:34.663646Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"rT/275DjRBPj/ifxLFFgQg/xXlAbCkNdD/ysWUScHWU0kR2XTpkjWw052zrKpbJbuTy7rup4h6bdn2Bf0y6YDg==","signature_status":"signed_v1","signed_at":"2026-07-05T09:24:34.664142Z","signed_message":"canonical_sha256_bytes"},"source_id":"2405.05439","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:bfd6208fcb65c6655aac0614e2d7a796cb7ee96da5b0275740f2a27f349d3100","sha256:963e00736882047f48e2e6564147342f09e5f3a4c28e5e7b559df26f684a0d0d"],"state_sha256":"8dd40b5543fea21b8c7d59e7096d8c5c13ab5c0288b9c78881e3dcec91ef5d56"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"xKzsAWdSBxfolP8EKiwnky1y0iVTM77C3SWArNrgiOKZgNkfr3s89kmM++r9x+sM4rkx7q6yjSEkZTGiaNJ0AA==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-03T19:43:41.464626Z","bundle_sha256":"0b2b01feec2df61510f010adbf8fcaf9097be4ff8687eece9983ab8632a0dd23"}}