{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:4SP77SIT3VWDW4WHQGQKWCTJXZ","short_pith_number":"pith:4SP77SIT","schema_version":"1.0","canonical_sha256":"e49fffc913dd6c3b72c781a0ab0a69be4293675defac070939f30f8d816f6d5a","source":{"kind":"arxiv","id":"2507.19468","version":1},"attestation_state":"computed","paper":{"title":"Back to the Features: DINO as a Foundation for Video World Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Basile Terver, Federico Baldassarre, Francisco Massa, Marc Szafraniec, Maximilian Seitzer, Patrick Labatut, Piotr Bojanowski, Vasil Khalidov, Yann LeCun","submitted_at":"2025-07-25T17:54:10Z","abstract_excerpt":"We present DINO-world, a powerful generalist video world model trained to predict future frames in the latent space of DINOv2. By leveraging a pre-trained image encoder and training a future predictor on a large-scale uncurated video dataset, DINO-world learns the temporal dynamics of diverse scenes, from driving and indoor scenes to simulated environments. We show that DINO-world outperforms previous models on a variety of video prediction benchmarks, e.g. segmentation and depth forecasting, and demonstrates strong understanding of intuitive physics. Furthermore, we show that it is possible t"},"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":"2507.19468","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2025-07-25T17:54:10Z","cross_cats_sorted":[],"title_canon_sha256":"654fbb6b0584009a750287dceb7dad732f5c407ec8cd616453ffd1ef9e8214ff","abstract_canon_sha256":"66b76aab849bb2c4a565893c3c20e2f68b90ccf364f5fd97e2476ba1dc5d615a"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:43:25.794711Z","signature_b64":"u2423307jQjwkTWEtNfu2DYEHTyEFdbgH3EhVOVwfDJWHqBiN3eKok0rK0777MnLWMWtW8+0wnMW+bDcLgZ0Dg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e49fffc913dd6c3b72c781a0ab0a69be4293675defac070939f30f8d816f6d5a","last_reissued_at":"2026-07-05T11:43:25.794231Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:43:25.794231Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Back to the Features: DINO as a Foundation for Video World Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Basile Terver, Federico Baldassarre, Francisco Massa, Marc Szafraniec, Maximilian Seitzer, Patrick Labatut, Piotr Bojanowski, Vasil Khalidov, Yann LeCun","submitted_at":"2025-07-25T17:54:10Z","abstract_excerpt":"We present DINO-world, a powerful generalist video world model trained to predict future frames in the latent space of DINOv2. By leveraging a pre-trained image encoder and training a future predictor on a large-scale uncurated video dataset, DINO-world learns the temporal dynamics of diverse scenes, from driving and indoor scenes to simulated environments. We show that DINO-world outperforms previous models on a variety of video prediction benchmarks, e.g. segmentation and depth forecasting, and demonstrates strong understanding of intuitive physics. Furthermore, we show that it is possible t"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2507.19468","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/2507.19468/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":"2507.19468","created_at":"2026-07-05T11:43:25.794296+00:00"},{"alias_kind":"arxiv_version","alias_value":"2507.19468v1","created_at":"2026-07-05T11:43:25.794296+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2507.19468","created_at":"2026-07-05T11:43:25.794296+00:00"},{"alias_kind":"pith_short_12","alias_value":"4SP77SIT3VWD","created_at":"2026-07-05T11:43:25.794296+00:00"},{"alias_kind":"pith_short_16","alias_value":"4SP77SIT3VWDW4WH","created_at":"2026-07-05T11:43:25.794296+00:00"},{"alias_kind":"pith_short_8","alias_value":"4SP77SIT","created_at":"2026-07-05T11:43:25.794296+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":11,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.18250","citing_title":"Future Dynamic 3D Reconstruction: A 3D World Model with Disentangled Ego-Motion","ref_index":2,"is_internal_anchor":false},{"citing_arxiv_id":"2606.12191","citing_title":"Agentic Environment Engineering for Large Language Models: A Survey of Environment Modeling, Synthesis, Evaluation, and Application","ref_index":214,"is_internal_anchor":false},{"citing_arxiv_id":"2606.10656","citing_title":"Envision4D: Envisioning Visual Futures via Feed-forward 4D Gaussian Splatting for Autonomous Driving","ref_index":52,"is_internal_anchor":false},{"citing_arxiv_id":"2605.25620","citing_title":"Back to Parsimonious Latents: Learning Task-Centric World Models from Visual Foundations","ref_index":3,"is_internal_anchor":false},{"citing_arxiv_id":"2512.24497","citing_title":"What Drives Success in Physical Planning with Joint-Embedding Predictive World Models?","ref_index":4,"is_internal_anchor":false},{"citing_arxiv_id":"2605.03245","citing_title":"Text-Conditional JEPA for Learning Semantically Rich Visual Representations","ref_index":2,"is_internal_anchor":false},{"citing_arxiv_id":"2605.06388","citing_title":"Reconstruction or Semantics? What Makes a Latent Space Useful for Robotic World Models","ref_index":8,"is_internal_anchor":false},{"citing_arxiv_id":"2605.02134","citing_title":"Video Generation with Predictive Latents","ref_index":3,"is_internal_anchor":false},{"citing_arxiv_id":"2604.11737","citing_title":"Learning Long-term Motion Embeddings for Efficient Kinematics Generation","ref_index":3,"is_internal_anchor":false},{"citing_arxiv_id":"2604.11707","citing_title":"Representations Before Pixels: Semantics-Guided Hierarchical Video Prediction","ref_index":6,"is_internal_anchor":false},{"citing_arxiv_id":"2605.07079","citing_title":"Learning Visual Feature-Based World Models via Residual Latent Action","ref_index":21,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/4SP77SIT3VWDW4WHQGQKWCTJXZ","json":"https://pith.science/pith/4SP77SIT3VWDW4WHQGQKWCTJXZ.json","graph_json":"https://pith.science/api/pith-number/4SP77SIT3VWDW4WHQGQKWCTJXZ/graph.json","events_json":"https://pith.science/api/pith-number/4SP77SIT3VWDW4WHQGQKWCTJXZ/events.json","paper":"https://pith.science/paper/4SP77SIT"},"agent_actions":{"view_html":"https://pith.science/pith/4SP77SIT3VWDW4WHQGQKWCTJXZ","download_json":"https://pith.science/pith/4SP77SIT3VWDW4WHQGQKWCTJXZ.json","view_paper":"https://pith.science/paper/4SP77SIT","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2507.19468&json=true","fetch_graph":"https://pith.science/api/pith-number/4SP77SIT3VWDW4WHQGQKWCTJXZ/graph.json","fetch_events":"https://pith.science/api/pith-number/4SP77SIT3VWDW4WHQGQKWCTJXZ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/4SP77SIT3VWDW4WHQGQKWCTJXZ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/4SP77SIT3VWDW4WHQGQKWCTJXZ/action/storage_attestation","attest_author":"https://pith.science/pith/4SP77SIT3VWDW4WHQGQKWCTJXZ/action/author_attestation","sign_citation":"https://pith.science/pith/4SP77SIT3VWDW4WHQGQKWCTJXZ/action/citation_signature","submit_replication":"https://pith.science/pith/4SP77SIT3VWDW4WHQGQKWCTJXZ/action/replication_record"}},"created_at":"2026-07-05T11:43:25.794296+00:00","updated_at":"2026-07-05T11:43:25.794296+00:00"}