{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:W7HUX3ZEAPLD5ZGDW3P7DG23WO","short_pith_number":"pith:W7HUX3ZE","schema_version":"1.0","canonical_sha256":"b7cf4bef2403d63ee4c3b6dff19b5bb392960124d2610daa45682ef6293801f8","source":{"kind":"arxiv","id":"2109.13396","version":1},"attestation_state":"computed","paper":{"title":"Bridge Data: Boosting Generalization of Robotic Skills with Cross-Domain Datasets","license":"http://creativecommons.org/licenses/by/4.0/","headline":"A shared multi-task multi-domain robot dataset doubles success rates for new tasks in new environments when added to just 50 demonstrations.","cross_cats":["cs.AI"],"primary_cat":"cs.RO","authors_text":"Bernadette Bucher, Chelsea Finn, Frederik Ebert, Georgios Georgakis, Karl Schmeckpeper, Kostas Daniilidis, Sergey Levine, Yanlai Yang","submitted_at":"2021-09-27T23:42:12Z","abstract_excerpt":"Robot learning holds the promise of learning policies that generalize broadly. However, such generalization requires sufficiently diverse datasets of the task of interest, which can be prohibitively expensive to collect. In other fields, such as computer vision, it is common to utilize shared, reusable datasets, such as ImageNet, to overcome this challenge, but this has proven difficult in robotics. In this paper, we ask: what would it take to enable practical data reuse in robotics for end-to-end skill learning? We hypothesize that the key is to use datasets with multiple tasks and multiple d"},"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":true,"formal_links_present":true},"canonical_record":{"source":{"id":"2109.13396","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.RO","submitted_at":"2021-09-27T23:42:12Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"77553db0d2eb8301edb63c8be3ac05770fc7f85bfd9e98af11142622cb1e35e7","abstract_canon_sha256":"d64a0c5131cb832077a8388f095a6403844f9cc0e1337b66058bbb17f28f9942"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T03:18:02.297284Z","signature_b64":"knl67scjyyj4tG/skItoF1INeMptQDVPch2tdEtSvb6dg/hiAg47nsvy1O4VUcbGyQjHWs6STFZsz2DYTQFVAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"b7cf4bef2403d63ee4c3b6dff19b5bb392960124d2610daa45682ef6293801f8","last_reissued_at":"2026-07-05T03:18:02.296857Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T03:18:02.296857Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Bridge Data: Boosting Generalization of Robotic Skills with Cross-Domain Datasets","license":"http://creativecommons.org/licenses/by/4.0/","headline":"A shared multi-task multi-domain robot dataset doubles success rates for new tasks in new environments when added to just 50 demonstrations.","cross_cats":["cs.AI"],"primary_cat":"cs.RO","authors_text":"Bernadette Bucher, Chelsea Finn, Frederik Ebert, Georgios Georgakis, Karl Schmeckpeper, Kostas Daniilidis, Sergey Levine, Yanlai Yang","submitted_at":"2021-09-27T23:42:12Z","abstract_excerpt":"Robot learning holds the promise of learning policies that generalize broadly. However, such generalization requires sufficiently diverse datasets of the task of interest, which can be prohibitively expensive to collect. In other fields, such as computer vision, it is common to utilize shared, reusable datasets, such as ImageNet, to overcome this challenge, but this has proven difficult in robotics. In this paper, we ask: what would it take to enable practical data reuse in robotics for end-to-end skill learning? We hypothesize that the key is to use datasets with multiple tasks and multiple d"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"jointly training with the proposed dataset and 50 demonstrations of a never-before-seen task in a new domain on average leads to a 2x improvement in success rate compared to using target domain data alone","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That the collected tasks and domains are representative enough that cross-domain data produces positive transfer rather than interference for arbitrary new tasks and environments.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"A large multi-task multi-domain robot dataset combined with 50 new demonstrations yields 2x higher success rates on never-before-seen tasks in new domains.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"A shared multi-task multi-domain robot dataset doubles success rates for new tasks in new environments when added to just 50 demonstrations.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"3a32b3447f97e7f0931f9c13d452d3333f56fac78f8b59aec215d9380a7dd31b"},"source":{"id":"2109.13396","kind":"arxiv","version":1},"verdict":{"id":"0f2e11a0-3517-4054-be91-9a38f96cd876","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-13T19:51:16.787575Z","strongest_claim":"jointly training with the proposed dataset and 50 demonstrations of a never-before-seen task in a new domain on average leads to a 2x improvement in success rate compared to using target domain data alone","one_line_summary":"A large multi-task multi-domain robot dataset combined with 50 new demonstrations yields 2x higher success rates on never-before-seen tasks in new domains.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That the collected tasks and domains are representative enough that cross-domain data produces positive transfer rather than interference for arbitrary new tasks and environments.","pith_extraction_headline":"A shared multi-task multi-domain robot dataset doubles success rates for new tasks in new environments when added to just 50 demonstrations."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2109.13396/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":28,"sample":[{"doi":"","year":2012,"title":"Imagenet classifica- tion with deep convolutional neural networks","work_id":"08ad5a97-cec1-4b1a-8d2f-af4f03c45e08","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2018,"title":"BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding","work_id":"ed240a10-5b19-406c-baa5-30803f465785","ref_index":2,"cited_arxiv_id":"1810.04805","is_internal_anchor":true},{"doi":"","year":2009,"title":"Imagenet: A large-scale hierarchical image database","work_id":"78bbc043-c6d8-4572-b5d6-54eaf5a89fb1","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2001,"title":"Gradient surgery for multi-task learning","work_id":"044e1624-6f75-4f7f-8757-624618e6015f","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2021,"title":"Mt-opt: Continuous multi-task robotic reinforcement learning at scale","work_id":"d4b61039-1d94-42db-8c6d-4c49c037e711","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":28,"snapshot_sha256":"350c1b7364023aef85c04f3a64a377c8bdf7a170f84e4c775899e0462a1d4f17","internal_anchors":2},"formal_canon":{"evidence_count":2,"snapshot_sha256":"eea2b2d019169268c3b3d6a9282b6cda74336b83d4913c4c1c8778c1d438f98f"},"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":"2109.13396","created_at":"2026-07-05T03:18:02.296921+00:00"},{"alias_kind":"arxiv_version","alias_value":"2109.13396v1","created_at":"2026-07-05T03:18:02.296921+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2109.13396","created_at":"2026-07-05T03:18:02.296921+00:00"},{"alias_kind":"pith_short_12","alias_value":"W7HUX3ZEAPLD","created_at":"2026-07-05T03:18:02.296921+00:00"},{"alias_kind":"pith_short_16","alias_value":"W7HUX3ZEAPLD5ZGD","created_at":"2026-07-05T03:18:02.296921+00:00"},{"alias_kind":"pith_short_8","alias_value":"W7HUX3ZE","created_at":"2026-07-05T03:18:02.296921+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":53,"internal_anchor_count":53,"sample":[{"citing_arxiv_id":"2606.22303","citing_title":"FlowDPG: Deterministic Policy Gradient on Flow Matching Policies for Real-World Manipulation","ref_index":6,"is_internal_anchor":true},{"citing_arxiv_id":"2606.19340","citing_title":"ZeroDex: Zero-Shot Long-Horizon Dexterous Manipulation via Multi-View 3D-Grounded VLM Reasoning","ref_index":10,"is_internal_anchor":true},{"citing_arxiv_id":"2606.18960","citing_title":"Mem-World: Memory-Augmented Action-Conditioned World Models for Persistent Robot Manipulation","ref_index":13,"is_internal_anchor":true},{"citing_arxiv_id":"2606.31836","citing_title":"RoboTacDex: A Dexterous Visual-Tactile-Action Dataset for Humanoid Manipulation","ref_index":2,"is_internal_anchor":true},{"citing_arxiv_id":"2605.29662","citing_title":"SAFE-Pruner: Semantic Attention-Guided Future-Aware Token Pruning for Efficient Vision-Language-Action Manipulation","ref_index":17,"is_internal_anchor":true},{"citing_arxiv_id":"2507.05331","citing_title":"A Careful Examination of Large Behavior Models for Multitask Dexterous Manipulation","ref_index":39,"is_internal_anchor":true},{"citing_arxiv_id":"2504.16054","citing_title":"$\\pi_{0.5}$: a Vision-Language-Action Model with Open-World Generalization","ref_index":25,"is_internal_anchor":true},{"citing_arxiv_id":"2605.22376","citing_title":"Target-Aligned Bellman Backup for Cross-domain Offline Reinforcement Learning","ref_index":5,"is_internal_anchor":true},{"citing_arxiv_id":"2605.17486","citing_title":"DyGRO-VLA: Cross-Task Scaling of Vision-Language-Action Models via Dynamic Grouped Residual Optimization","ref_index":35,"is_internal_anchor":true},{"citing_arxiv_id":"2511.01770","citing_title":"Lightweight Learning from Actuation-Space Demonstrations via Flow Matching for Whole-Body Soft Robotic Grasping","ref_index":22,"is_internal_anchor":true},{"citing_arxiv_id":"2511.17441","citing_title":"RoboCOIN: An Open-Sourced Bimanual Robotic Data Collection for Integrated Manipulation","ref_index":10,"is_internal_anchor":true},{"citing_arxiv_id":"2302.11550","citing_title":"Scaling Robot Learning with Semantically Imagined Experience","ref_index":31,"is_internal_anchor":true},{"citing_arxiv_id":"2507.01925","citing_title":"A Survey on Vision-Language-Action Models: An Action Tokenization Perspective","ref_index":205,"is_internal_anchor":true},{"citing_arxiv_id":"2310.17596","citing_title":"MimicGen: A Data Generation System for Scalable Robot Learning using Human Demonstrations","ref_index":6,"is_internal_anchor":true},{"citing_arxiv_id":"2507.15493","citing_title":"GR-3 Technical Report","ref_index":21,"is_internal_anchor":true},{"citing_arxiv_id":"2512.01773","citing_title":"IGen: Scalable Data Generation for Robot Learning from Open-World Images","ref_index":16,"is_internal_anchor":true},{"citing_arxiv_id":"2512.15840","citing_title":"Large Video Planner Enables Generalizable Robot Control","ref_index":27,"is_internal_anchor":true},{"citing_arxiv_id":"2504.19854","citing_title":"NORA: A Small Open-Sourced Generalist Vision Language Action Model for Embodied Tasks","ref_index":6,"is_internal_anchor":true},{"citing_arxiv_id":"2507.04447","citing_title":"DreamVLA: A Vision-Language-Action Model Dreamed with Comprehensive World Knowledge","ref_index":81,"is_internal_anchor":true},{"citing_arxiv_id":"2601.07060","citing_title":"PALM: Progress-Aware Policy Learning via Affordance Reasoning for Long-Horizon Robotic Manipulation","ref_index":27,"is_internal_anchor":true},{"citing_arxiv_id":"2505.18719","citing_title":"VLA-RL: Towards Masterful and General Robotic Manipulation with Scalable Reinforcement Learning","ref_index":18,"is_internal_anchor":true},{"citing_arxiv_id":"2503.22020","citing_title":"CoT-VLA: Visual Chain-of-Thought Reasoning for Vision-Language-Action Models","ref_index":16,"is_internal_anchor":true},{"citing_arxiv_id":"2310.06114","citing_title":"Learning Interactive Real-World Simulators","ref_index":223,"is_internal_anchor":true},{"citing_arxiv_id":"2503.10631","citing_title":"HybridVLA: Collaborative Diffusion and Autoregression in a Unified Vision-Language-Action Model","ref_index":85,"is_internal_anchor":true},{"citing_arxiv_id":"2507.23682","citing_title":"villa-X: Enhancing Latent Action Modeling in Vision-Language-Action Models","ref_index":17,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":2,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/W7HUX3ZEAPLD5ZGDW3P7DG23WO","json":"https://pith.science/pith/W7HUX3ZEAPLD5ZGDW3P7DG23WO.json","graph_json":"https://pith.science/api/pith-number/W7HUX3ZEAPLD5ZGDW3P7DG23WO/graph.json","events_json":"https://pith.science/api/pith-number/W7HUX3ZEAPLD5ZGDW3P7DG23WO/events.json","paper":"https://pith.science/paper/W7HUX3ZE"},"agent_actions":{"view_html":"https://pith.science/pith/W7HUX3ZEAPLD5ZGDW3P7DG23WO","download_json":"https://pith.science/pith/W7HUX3ZEAPLD5ZGDW3P7DG23WO.json","view_paper":"https://pith.science/paper/W7HUX3ZE","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2109.13396&json=true","fetch_graph":"https://pith.science/api/pith-number/W7HUX3ZEAPLD5ZGDW3P7DG23WO/graph.json","fetch_events":"https://pith.science/api/pith-number/W7HUX3ZEAPLD5ZGDW3P7DG23WO/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/W7HUX3ZEAPLD5ZGDW3P7DG23WO/action/timestamp_anchor","attest_storage":"https://pith.science/pith/W7HUX3ZEAPLD5ZGDW3P7DG23WO/action/storage_attestation","attest_author":"https://pith.science/pith/W7HUX3ZEAPLD5ZGDW3P7DG23WO/action/author_attestation","sign_citation":"https://pith.science/pith/W7HUX3ZEAPLD5ZGDW3P7DG23WO/action/citation_signature","submit_replication":"https://pith.science/pith/W7HUX3ZEAPLD5ZGDW3P7DG23WO/action/replication_record"}},"created_at":"2026-07-05T03:18:02.296921+00:00","updated_at":"2026-07-05T03:18:02.296921+00:00"}