{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:2BSMLSOVTEKLN32W3OIYFH3XEC","short_pith_number":"pith:2BSMLSOV","schema_version":"1.0","canonical_sha256":"d064c5c9d59914b6ef56db91829f77209c409371dc46304ea119933d61f29a92","source":{"kind":"arxiv","id":"2406.18158","version":2},"attestation_state":"computed","paper":{"title":"3D-MVP: 3D Multiview Pretraining for Robotic Manipulation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CV"],"primary_cat":"cs.RO","authors_text":"Ankit Goyal, David F. Fouhey, Dieter Fox, Kaichun Mo, Shengyi Qian, Valts Blukis","submitted_at":"2024-06-26T08:17:59Z","abstract_excerpt":"Recent works have shown that visual pretraining on egocentric datasets using masked autoencoders (MAE) can improve generalization for downstream robotics tasks. However, these approaches pretrain only on 2D images, while many robotics applications require 3D scene understanding. In this work, we propose 3D-MVP, a novel approach for 3D Multi-View Pretraining using masked autoencoders. We leverage Robotic View Transformer (RVT), which uses a multi-view transformer to understand the 3D scene and predict gripper pose actions. We split RVT's multi-view transformer into visual encoder and action dec"},"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":"2406.18158","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.RO","submitted_at":"2024-06-26T08:17:59Z","cross_cats_sorted":["cs.CV"],"title_canon_sha256":"7477f7ec9da094fb6374632132178aac8d1f1a746df34252e337e7439515b079","abstract_canon_sha256":"6c46359829dc70ddd2057f322c7a299e9e9d55bf60b39d2f1411e9eeee54a63f"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:37:27.865979Z","signature_b64":"Km/pnNfvrdZvAn1BYV/0zD/k3jz+4MP2HVxraQ1abx6TfzWyl4fRxrEgIcSXs/bf35Y4b2T2DRKGTc2LR9gRAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"d064c5c9d59914b6ef56db91829f77209c409371dc46304ea119933d61f29a92","last_reissued_at":"2026-07-05T10:37:27.865507Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:37:27.865507Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"3D-MVP: 3D Multiview Pretraining for Robotic Manipulation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CV"],"primary_cat":"cs.RO","authors_text":"Ankit Goyal, David F. Fouhey, Dieter Fox, Kaichun Mo, Shengyi Qian, Valts Blukis","submitted_at":"2024-06-26T08:17:59Z","abstract_excerpt":"Recent works have shown that visual pretraining on egocentric datasets using masked autoencoders (MAE) can improve generalization for downstream robotics tasks. However, these approaches pretrain only on 2D images, while many robotics applications require 3D scene understanding. In this work, we propose 3D-MVP, a novel approach for 3D Multi-View Pretraining using masked autoencoders. We leverage Robotic View Transformer (RVT), which uses a multi-view transformer to understand the 3D scene and predict gripper pose actions. We split RVT's multi-view transformer into visual encoder and action dec"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2406.18158","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/2406.18158/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":"2406.18158","created_at":"2026-07-05T10:37:27.865565+00:00"},{"alias_kind":"arxiv_version","alias_value":"2406.18158v2","created_at":"2026-07-05T10:37:27.865565+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2406.18158","created_at":"2026-07-05T10:37:27.865565+00:00"},{"alias_kind":"pith_short_12","alias_value":"2BSMLSOVTEKL","created_at":"2026-07-05T10:37:27.865565+00:00"},{"alias_kind":"pith_short_16","alias_value":"2BSMLSOVTEKLN32W","created_at":"2026-07-05T10:37:27.865565+00:00"},{"alias_kind":"pith_short_8","alias_value":"2BSMLSOV","created_at":"2026-07-05T10:37:27.865565+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":4,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2506.19579","citing_title":"Fake or Real, Can Robots Tell? Evaluating VLM Robustness to Domain Shift in Single-View Robotic Scene Understanding","ref_index":18,"is_internal_anchor":false},{"citing_arxiv_id":"2604.03878","citing_title":"Learning 3D Reconstruction with Priors in Test Time","ref_index":34,"is_internal_anchor":false},{"citing_arxiv_id":"2604.09824","citing_title":"ProGAL-VLA: Grounded Alignment through Prospective Reasoning in Vision-Language-Action Models","ref_index":27,"is_internal_anchor":false},{"citing_arxiv_id":"2604.15281","citing_title":"R3D: Revisiting 3D Policy Learning","ref_index":34,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/2BSMLSOVTEKLN32W3OIYFH3XEC","json":"https://pith.science/pith/2BSMLSOVTEKLN32W3OIYFH3XEC.json","graph_json":"https://pith.science/api/pith-number/2BSMLSOVTEKLN32W3OIYFH3XEC/graph.json","events_json":"https://pith.science/api/pith-number/2BSMLSOVTEKLN32W3OIYFH3XEC/events.json","paper":"https://pith.science/paper/2BSMLSOV"},"agent_actions":{"view_html":"https://pith.science/pith/2BSMLSOVTEKLN32W3OIYFH3XEC","download_json":"https://pith.science/pith/2BSMLSOVTEKLN32W3OIYFH3XEC.json","view_paper":"https://pith.science/paper/2BSMLSOV","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2406.18158&json=true","fetch_graph":"https://pith.science/api/pith-number/2BSMLSOVTEKLN32W3OIYFH3XEC/graph.json","fetch_events":"https://pith.science/api/pith-number/2BSMLSOVTEKLN32W3OIYFH3XEC/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/2BSMLSOVTEKLN32W3OIYFH3XEC/action/timestamp_anchor","attest_storage":"https://pith.science/pith/2BSMLSOVTEKLN32W3OIYFH3XEC/action/storage_attestation","attest_author":"https://pith.science/pith/2BSMLSOVTEKLN32W3OIYFH3XEC/action/author_attestation","sign_citation":"https://pith.science/pith/2BSMLSOVTEKLN32W3OIYFH3XEC/action/citation_signature","submit_replication":"https://pith.science/pith/2BSMLSOVTEKLN32W3OIYFH3XEC/action/replication_record"}},"created_at":"2026-07-05T10:37:27.865565+00:00","updated_at":"2026-07-05T10:37:27.865565+00:00"}