{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:YWHVPC55YXHETDN6FYVUZ23CVR","short_pith_number":"pith:YWHVPC55","schema_version":"1.0","canonical_sha256":"c58f578bbdc5ce498dbe2e2b4ceb62ac7b19f3532a732e1d65946c673aabcd70","source":{"kind":"arxiv","id":"2503.14526","version":1},"attestation_state":"computed","paper":{"title":"ReBot: Scaling Robot Learning with Real-to-Sim-to-Real Robotic Video Synthesis","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.GR","cs.RO"],"primary_cat":"cs.CV","authors_text":"Daniel Szafir, Gedas Bertasius, Kaiyuan Zheng, Mingyu Ding, Xinghao Zhu, Yue Yang, Yu Fang","submitted_at":"2025-03-15T16:47:25Z","abstract_excerpt":"Vision-language-action (VLA) models present a promising paradigm by training policies directly on real robot datasets like Open X-Embodiment. However, the high cost of real-world data collection hinders further data scaling, thereby restricting the generalizability of VLAs. In this paper, we introduce ReBot, a novel real-to-sim-to-real approach for scaling real robot datasets and adapting VLA models to target domains, which is the last-mile deployment challenge in robot manipulation. Specifically, ReBot replays real-world robot trajectories in simulation to diversify manipulated objects (real-"},"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":"2503.14526","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2025-03-15T16:47:25Z","cross_cats_sorted":["cs.GR","cs.RO"],"title_canon_sha256":"7fe20705a49f59b73e324bd9d07a162154916629fc894cee28f58504c940b9cb","abstract_canon_sha256":"d8789fd0f149dc575ff5841d6ff6fd4cee7b7ebae995808106ff688edd1fd7fa"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:34:19.345195Z","signature_b64":"2TQptfyTK3tz5eNqbaLA7skzug7XyQwfLUFWVA54QKCApFF9ENj9wFWDcMuMFKxajLpl3DJT4296K73hiO+OBA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"c58f578bbdc5ce498dbe2e2b4ceb62ac7b19f3532a732e1d65946c673aabcd70","last_reissued_at":"2026-07-05T10:34:19.344728Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:34:19.344728Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"ReBot: Scaling Robot Learning with Real-to-Sim-to-Real Robotic Video Synthesis","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.GR","cs.RO"],"primary_cat":"cs.CV","authors_text":"Daniel Szafir, Gedas Bertasius, Kaiyuan Zheng, Mingyu Ding, Xinghao Zhu, Yue Yang, Yu Fang","submitted_at":"2025-03-15T16:47:25Z","abstract_excerpt":"Vision-language-action (VLA) models present a promising paradigm by training policies directly on real robot datasets like Open X-Embodiment. However, the high cost of real-world data collection hinders further data scaling, thereby restricting the generalizability of VLAs. In this paper, we introduce ReBot, a novel real-to-sim-to-real approach for scaling real robot datasets and adapting VLA models to target domains, which is the last-mile deployment challenge in robot manipulation. Specifically, ReBot replays real-world robot trajectories in simulation to diversify manipulated objects (real-"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2503.14526","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/2503.14526/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":"2503.14526","created_at":"2026-07-05T10:34:19.344789+00:00"},{"alias_kind":"arxiv_version","alias_value":"2503.14526v1","created_at":"2026-07-05T10:34:19.344789+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2503.14526","created_at":"2026-07-05T10:34:19.344789+00:00"},{"alias_kind":"pith_short_12","alias_value":"YWHVPC55YXHE","created_at":"2026-07-05T10:34:19.344789+00:00"},{"alias_kind":"pith_short_16","alias_value":"YWHVPC55YXHETDN6","created_at":"2026-07-05T10:34:19.344789+00:00"},{"alias_kind":"pith_short_8","alias_value":"YWHVPC55","created_at":"2026-07-05T10:34:19.344789+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":3,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2607.06699","citing_title":"RoboSnap: One-Shot Real-to-Sim Scene Generation for Generalizable Robot Learning and Evaluation","ref_index":38,"is_internal_anchor":true},{"citing_arxiv_id":"2606.16776","citing_title":"JoyAI-Sim: A Simulation-Enabled Interconversion Toolchain for the Embodied Data Pyramid","ref_index":12,"is_internal_anchor":false},{"citing_arxiv_id":"2605.15836","citing_title":"GAP: Geometric Anchor Pre-training for Data-Efficient Visuomotor Learning of Manipulation Tasks","ref_index":23,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/YWHVPC55YXHETDN6FYVUZ23CVR","json":"https://pith.science/pith/YWHVPC55YXHETDN6FYVUZ23CVR.json","graph_json":"https://pith.science/api/pith-number/YWHVPC55YXHETDN6FYVUZ23CVR/graph.json","events_json":"https://pith.science/api/pith-number/YWHVPC55YXHETDN6FYVUZ23CVR/events.json","paper":"https://pith.science/paper/YWHVPC55"},"agent_actions":{"view_html":"https://pith.science/pith/YWHVPC55YXHETDN6FYVUZ23CVR","download_json":"https://pith.science/pith/YWHVPC55YXHETDN6FYVUZ23CVR.json","view_paper":"https://pith.science/paper/YWHVPC55","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2503.14526&json=true","fetch_graph":"https://pith.science/api/pith-number/YWHVPC55YXHETDN6FYVUZ23CVR/graph.json","fetch_events":"https://pith.science/api/pith-number/YWHVPC55YXHETDN6FYVUZ23CVR/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/YWHVPC55YXHETDN6FYVUZ23CVR/action/timestamp_anchor","attest_storage":"https://pith.science/pith/YWHVPC55YXHETDN6FYVUZ23CVR/action/storage_attestation","attest_author":"https://pith.science/pith/YWHVPC55YXHETDN6FYVUZ23CVR/action/author_attestation","sign_citation":"https://pith.science/pith/YWHVPC55YXHETDN6FYVUZ23CVR/action/citation_signature","submit_replication":"https://pith.science/pith/YWHVPC55YXHETDN6FYVUZ23CVR/action/replication_record"}},"created_at":"2026-07-05T10:34:19.344789+00:00","updated_at":"2026-07-05T10:34:19.344789+00:00"}