{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:WTEPQDWKB5OIJAEZQW4YETSSJC","short_pith_number":"pith:WTEPQDWK","schema_version":"1.0","canonical_sha256":"b4c8f80eca0f5c84809985b9824e5248ad16aba97b4dba5ae6e3aab69a0e3ced","source":{"kind":"arxiv","id":"2307.14619","version":5},"attestation_state":"computed","paper":{"title":"Provable Guarantees for Generative Behavior Cloning: Bridging Low-Level Stability and High-Level Behavior","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["math.ST","stat.ML","stat.TH"],"primary_cat":"cs.LG","authors_text":"Adam Block, Ali Jadbabaie, Daniel Pfrommer, Max Simchowitz, Russ Tedrake","submitted_at":"2023-07-27T04:27:26Z","abstract_excerpt":"We propose a theoretical framework for studying behavior cloning of complex expert demonstrations using generative modeling. Our framework invokes low-level controllers - either learned or implicit in position-command control - to stabilize imitation around expert demonstrations. We show that with (a) a suitable low-level stability guarantee and (b) a powerful enough generative model as our imitation learner, pure supervised behavior cloning can generate trajectories matching the per-time step distribution of essentially arbitrary expert trajectories in an optimal transport cost. Our analysis "},"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":"2307.14619","kind":"arxiv","version":5},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2023-07-27T04:27:26Z","cross_cats_sorted":["math.ST","stat.ML","stat.TH"],"title_canon_sha256":"782e17b82a288ad1afbf4a689abcd652e0b7da12952fe54b598ad48ea17b8ead","abstract_canon_sha256":"37e2853ae8e1d963a37270901021399313b53bf9dfa069c99405f3d503d014e5"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T07:04:18.348038Z","signature_b64":"ucBMtrgMBM/VXd//0a4TT1hcwENrJxyNkT8ma1jC6c1ilIT8RT9/7abA4oshTt8wK6QvQszbd/hBRBFycZ69BQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"b4c8f80eca0f5c84809985b9824e5248ad16aba97b4dba5ae6e3aab69a0e3ced","last_reissued_at":"2026-07-05T07:04:18.347464Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T07:04:18.347464Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Provable Guarantees for Generative Behavior Cloning: Bridging Low-Level Stability and High-Level Behavior","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["math.ST","stat.ML","stat.TH"],"primary_cat":"cs.LG","authors_text":"Adam Block, Ali Jadbabaie, Daniel Pfrommer, Max Simchowitz, Russ Tedrake","submitted_at":"2023-07-27T04:27:26Z","abstract_excerpt":"We propose a theoretical framework for studying behavior cloning of complex expert demonstrations using generative modeling. Our framework invokes low-level controllers - either learned or implicit in position-command control - to stabilize imitation around expert demonstrations. We show that with (a) a suitable low-level stability guarantee and (b) a powerful enough generative model as our imitation learner, pure supervised behavior cloning can generate trajectories matching the per-time step distribution of essentially arbitrary expert trajectories in an optimal transport cost. Our analysis "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2307.14619","kind":"arxiv","version":5},"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/2307.14619/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":"2307.14619","created_at":"2026-07-05T07:04:18.347539+00:00"},{"alias_kind":"arxiv_version","alias_value":"2307.14619v5","created_at":"2026-07-05T07:04:18.347539+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2307.14619","created_at":"2026-07-05T07:04:18.347539+00:00"},{"alias_kind":"pith_short_12","alias_value":"WTEPQDWKB5OI","created_at":"2026-07-05T07:04:18.347539+00:00"},{"alias_kind":"pith_short_16","alias_value":"WTEPQDWKB5OIJAEZ","created_at":"2026-07-05T07:04:18.347539+00:00"},{"alias_kind":"pith_short_8","alias_value":"WTEPQDWK","created_at":"2026-07-05T07:04:18.347539+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.16520","citing_title":"Global Convergence of Sampling-Based Nonconvex Optimization through Diffusion-Style Smoothing","ref_index":29,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/WTEPQDWKB5OIJAEZQW4YETSSJC","json":"https://pith.science/pith/WTEPQDWKB5OIJAEZQW4YETSSJC.json","graph_json":"https://pith.science/api/pith-number/WTEPQDWKB5OIJAEZQW4YETSSJC/graph.json","events_json":"https://pith.science/api/pith-number/WTEPQDWKB5OIJAEZQW4YETSSJC/events.json","paper":"https://pith.science/paper/WTEPQDWK"},"agent_actions":{"view_html":"https://pith.science/pith/WTEPQDWKB5OIJAEZQW4YETSSJC","download_json":"https://pith.science/pith/WTEPQDWKB5OIJAEZQW4YETSSJC.json","view_paper":"https://pith.science/paper/WTEPQDWK","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2307.14619&json=true","fetch_graph":"https://pith.science/api/pith-number/WTEPQDWKB5OIJAEZQW4YETSSJC/graph.json","fetch_events":"https://pith.science/api/pith-number/WTEPQDWKB5OIJAEZQW4YETSSJC/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/WTEPQDWKB5OIJAEZQW4YETSSJC/action/timestamp_anchor","attest_storage":"https://pith.science/pith/WTEPQDWKB5OIJAEZQW4YETSSJC/action/storage_attestation","attest_author":"https://pith.science/pith/WTEPQDWKB5OIJAEZQW4YETSSJC/action/author_attestation","sign_citation":"https://pith.science/pith/WTEPQDWKB5OIJAEZQW4YETSSJC/action/citation_signature","submit_replication":"https://pith.science/pith/WTEPQDWKB5OIJAEZQW4YETSSJC/action/replication_record"}},"created_at":"2026-07-05T07:04:18.347539+00:00","updated_at":"2026-07-05T07:04:18.347539+00:00"}