{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:57MOY35RKSWDA3PUOXJE5ICEAW","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":"e58d581b062dd34ba2e4bdac3fef6fec60b35cb830b49b5caf1d525671d436f7","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.RO","submitted_at":"2024-11-07T18:54:37Z","title_canon_sha256":"40b94add53985efde7ec3e35e5d2747436cc6d0ef9d700f818ce229c4e7b9abb"},"schema_version":"1.0","source":{"id":"2411.04983","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2411.04983","created_at":"2026-05-17T23:38:13Z"},{"alias_kind":"arxiv_version","alias_value":"2411.04983v2","created_at":"2026-05-17T23:38:13Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2411.04983","created_at":"2026-05-17T23:38:13Z"},{"alias_kind":"pith_short_12","alias_value":"57MOY35RKSWD","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_16","alias_value":"57MOY35RKSWDA3PU","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_8","alias_value":"57MOY35R","created_at":"2026-05-18T12:33:37Z"}],"graph_snapshots":[{"event_id":"sha256:ebce2753c3997dbcc4450b9d8e8c5dd7f01d0f934b7bfca2f211478f66b8c968","target":"graph","created_at":"2026-05-17T23:38:13Z","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":4,"items":[{"attestation":"unclaimed","claim_id":"C1","kind":"strongest_claim","source":"verdict.strongest_claim","status":"machine_extracted","text":"DINO-WM achieves zero-shot behavioral solutions at test time on six environments without expert demonstrations, reward modeling, or pre-learned inverse models, outperforming prior state-of-the-art work across diverse task families."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"That predicting future DINOv2 patch features alone captures sufficient dynamics information to enable reliable planning without visual reconstruction or task-specific components."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"DINO-WM builds world models on pre-trained DINOv2 features to enable zero-shot planning from offline data without rewards or demonstrations."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"DINO-WM uses pre-trained DINOv2 patch features to build world models that support zero-shot planning from offline data."}],"snapshot_sha256":"6db11558a2ac2cfa3ee8711725c5f65e23582492548685ec38060de65901eeef"},"formal_canon":{"evidence_count":2,"snapshot_sha256":"df7eca548065fee867ef9495de757f98066fad6fe5dcac2a027416b1545f576c"},"paper":{"abstract_excerpt":"The ability to predict future outcomes given control actions is fundamental for physical reasoning. However, such predictive models, often called world models, remains challenging to learn and are typically developed for task-specific solutions with online policy learning. To unlock world models' true potential, we argue that they should 1) be trainable on offline, pre-collected trajectories, 2) support test-time behavior optimization, and 3) facilitate task-agnostic reasoning. To this end, we present DINO World Model (DINO-WM), a new method to model visual dynamics without reconstructing the ","authors_text":"Gaoyue Zhou, Hengkai Pan, Lerrel Pinto, Yann LeCun","cross_cats":["cs.AI"],"headline":"DINO-WM uses pre-trained DINOv2 patch features to build world models that support zero-shot planning from offline data.","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.RO","submitted_at":"2024-11-07T18:54:37Z","title":"DINO-WM: World Models on Pre-trained Visual Features enable Zero-shot Planning"},"references":{"count":131,"internal_anchors":27,"resolved_work":131,"sample":[{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":1,"title":"Legged locomotion in challenging terrains using egocentric vision, 2022","work_id":"205d3229-9a4d-4506-8604-df6fa05bdde0","year":2022},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":2,"title":"Self-supervised learning from images with a joint-embedding predictive architecture","work_id":"f102eaff-c9af-4951-acb5-35e7356393be","year":2023},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":3,"title":"Nonlinear and adaptive control with applications, volume 187","work_id":"c5bc28bc-c8f6-4af9-a97f-04c76ed022a8","year":2008},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":4,"title":"V- JEPA : Latent video prediction for visual representation learning, 2024","work_id":"42a222c5-454c-48a6-b512-a7383176ac86","year":2024},{"cited_arxiv_id":"2307.15818","doi":"","is_internal_anchor":true,"ref_index":5,"title":"RT-2: Vision-Language-Action Models Transfer Web Knowledge to Robotic Control","work_id":"ff438a8a-8003-4fae-9131-acd418b3597b","year":2023}],"snapshot_sha256":"e16228a6869cc6f36edf1c4edfcb286399aa4c4123e66128ea91fb955b21680b"},"source":{"id":"2411.04983","kind":"arxiv","version":2},"verdict":{"created_at":"2026-05-17T16:00:46.891348Z","id":"12c5d9cf-9f77-4a05-9637-ab014ad26b52","model_set":{"reader":"grok-4.3"},"one_line_summary":"DINO-WM builds world models on pre-trained DINOv2 features to enable zero-shot planning from offline data without rewards or demonstrations.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"DINO-WM uses pre-trained DINOv2 patch features to build world models that support zero-shot planning from offline data.","strongest_claim":"DINO-WM achieves zero-shot behavioral solutions at test time on six environments without expert demonstrations, reward modeling, or pre-learned inverse models, outperforming prior state-of-the-art work across diverse task families.","weakest_assumption":"That predicting future DINOv2 patch features alone captures sufficient dynamics information to enable reliable planning without visual reconstruction or task-specific components."}},"verdict_id":"12c5d9cf-9f77-4a05-9637-ab014ad26b52"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:75784af86d0668c8f9f153608cdc8179a344ce94b6dfb5cf053d5f3c5571a525","target":"record","created_at":"2026-05-17T23:38:13Z","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":"e58d581b062dd34ba2e4bdac3fef6fec60b35cb830b49b5caf1d525671d436f7","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.RO","submitted_at":"2024-11-07T18:54:37Z","title_canon_sha256":"40b94add53985efde7ec3e35e5d2747436cc6d0ef9d700f818ce229c4e7b9abb"},"schema_version":"1.0","source":{"id":"2411.04983","kind":"arxiv","version":2}},"canonical_sha256":"efd8ec6fb154ac306df475d24ea044059ae2ad452412b6a444c84a360b52da61","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"efd8ec6fb154ac306df475d24ea044059ae2ad452412b6a444c84a360b52da61","first_computed_at":"2026-05-17T23:38:13.635938Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-17T23:38:13.635938Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"AQBCt3y22UEELqNGb/2ITSEAezzYTHuMvhzPLNFrRv7/Xx1dbIHcOjDY4S/aOHYjZli2RdxC6nwopr8Nm0TqAg==","signature_status":"signed_v1","signed_at":"2026-05-17T23:38:13.636570Z","signed_message":"canonical_sha256_bytes"},"source_id":"2411.04983","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:75784af86d0668c8f9f153608cdc8179a344ce94b6dfb5cf053d5f3c5571a525","sha256:ebce2753c3997dbcc4450b9d8e8c5dd7f01d0f934b7bfca2f211478f66b8c968"],"state_sha256":"ea5c710dd1173c074de841b6beb197998f5b23f8367902cbe313653fd16f4dd9"}