{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:BXELDHNB7GLCBRW66BR2LTWIJA","short_pith_number":"pith:BXELDHNB","schema_version":"1.0","canonical_sha256":"0dc8b19da1f99620c6def063a5cec84807020b5dd6d9c3ece046aa1f89162885","source":{"kind":"arxiv","id":"2506.05284","version":1},"attestation_state":"computed","paper":{"title":"Video World Models with Long-term Spatial Memory","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Dahua Lin, Gordon Wetzstein, Ryan Po, Shuai Yang, Tong Wu, Yinghao Xu, Ziwei Liu","submitted_at":"2025-06-05T17:42:34Z","abstract_excerpt":"Emerging world models autoregressively generate video frames in response to actions, such as camera movements and text prompts, among other control signals. Due to limited temporal context window sizes, these models often struggle to maintain scene consistency during revisits, leading to severe forgetting of previously generated environments. Inspired by the mechanisms of human memory, we introduce a novel framework to enhancing long-term consistency of video world models through a geometry-grounded long-term spatial memory. Our framework includes mechanisms to store and retrieve information f"},"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":"2506.05284","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2025-06-05T17:42:34Z","cross_cats_sorted":[],"title_canon_sha256":"df3f97e3f38d939190388d42971a0b44f2a9d567e833c405055fbf4ba138cedc","abstract_canon_sha256":"d91e55c816fe709bb598b4ccadf351dfb003778042ae53939fa4e80d66ac597f"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:16:45.169574Z","signature_b64":"JAEeuTRBMoWtfJ8tPhIezJtWga5Dos8ELowh6ZzMKZhBz98jFhGg+3u4+TBPZ/9vvwk0jc3axVTx3lkkJHYiAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"0dc8b19da1f99620c6def063a5cec84807020b5dd6d9c3ece046aa1f89162885","last_reissued_at":"2026-07-05T11:16:45.169075Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:16:45.169075Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Video World Models with Long-term Spatial Memory","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Dahua Lin, Gordon Wetzstein, Ryan Po, Shuai Yang, Tong Wu, Yinghao Xu, Ziwei Liu","submitted_at":"2025-06-05T17:42:34Z","abstract_excerpt":"Emerging world models autoregressively generate video frames in response to actions, such as camera movements and text prompts, among other control signals. Due to limited temporal context window sizes, these models often struggle to maintain scene consistency during revisits, leading to severe forgetting of previously generated environments. Inspired by the mechanisms of human memory, we introduce a novel framework to enhancing long-term consistency of video world models through a geometry-grounded long-term spatial memory. Our framework includes mechanisms to store and retrieve information f"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2506.05284","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/2506.05284/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":"2506.05284","created_at":"2026-07-05T11:16:45.169136+00:00"},{"alias_kind":"arxiv_version","alias_value":"2506.05284v1","created_at":"2026-07-05T11:16:45.169136+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2506.05284","created_at":"2026-07-05T11:16:45.169136+00:00"},{"alias_kind":"pith_short_12","alias_value":"BXELDHNB7GLC","created_at":"2026-07-05T11:16:45.169136+00:00"},{"alias_kind":"pith_short_16","alias_value":"BXELDHNB7GLCBRW6","created_at":"2026-07-05T11:16:45.169136+00:00"},{"alias_kind":"pith_short_8","alias_value":"BXELDHNB","created_at":"2026-07-05T11:16:45.169136+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":31,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.16449","citing_title":"PermaVid: Consistent Video Generation Across Edits via Disentangled Context Memory","ref_index":4,"is_internal_anchor":false},{"citing_arxiv_id":"2606.11129","citing_title":"WorldOlympiad: Can Your World Model Survive a Triathlon?","ref_index":42,"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":44,"is_internal_anchor":false},{"citing_arxiv_id":"2606.09828","citing_title":"Latent Spatial Memory for Video World Models","ref_index":18,"is_internal_anchor":false},{"citing_arxiv_id":"2606.09803","citing_title":"Echo-Memory: A Controlled Study of Memory in Action World Models","ref_index":57,"is_internal_anchor":false},{"citing_arxiv_id":"2607.00832","citing_title":"Pano2World: End-to-End 3D Generation via Unified Multi-View Sequences","ref_index":19,"is_internal_anchor":false},{"citing_arxiv_id":"2607.01202","citing_title":"World from Motion: Generative Dynamic Gaussian Reconstruction from Monocular Video","ref_index":59,"is_internal_anchor":false},{"citing_arxiv_id":"2606.02753","citing_title":"MetaWorld: Scaling Multi-Agent Video World Model from Single-view Video Data","ref_index":31,"is_internal_anchor":false},{"citing_arxiv_id":"2606.02436","citing_title":"Geometry-Aware Implicit Memory for Video World Models","ref_index":55,"is_internal_anchor":false},{"citing_arxiv_id":"2605.30855","citing_title":"Robust Dreamer: Deviation-Aware Latent Gaussian Memory for Action-Controlled AR Video Generation","ref_index":63,"is_internal_anchor":false},{"citing_arxiv_id":"2606.31734","citing_title":"MemLearner: Learning to Query Context memory for Video World Models","ref_index":56,"is_internal_anchor":false},{"citing_arxiv_id":"2605.26316","citing_title":"E$^3$C: Video Generation with 3D Environmental Memory and Ego-Exo Human Pose Control","ref_index":64,"is_internal_anchor":false},{"citing_arxiv_id":"2606.27964","citing_title":"Directing the World: Fast Autoregressive Video Generation with Compositional Human-Camera Control","ref_index":61,"is_internal_anchor":false},{"citing_arxiv_id":"2605.27589","citing_title":"What-If World: A Causal Benchmark for General World Models in Embodied Scenarios","ref_index":69,"is_internal_anchor":false},{"citing_arxiv_id":"2605.31336","citing_title":"DecMem: Towards Minute-Long Consistent World Generation with Decoupled Memory","ref_index":41,"is_internal_anchor":false},{"citing_arxiv_id":"2604.24764","citing_title":"World-R1: Reinforcing 3D Constraints for Text-to-Video Generation","ref_index":9,"is_internal_anchor":false},{"citing_arxiv_id":"2605.22814","citing_title":"Remember to be Curious: Episodic Context and Persistent Worlds for 3D Exploration","ref_index":67,"is_internal_anchor":false},{"citing_arxiv_id":"2604.24764","citing_title":"World-R1: Reinforcing 3D Constraints for Text-to-Video Generation","ref_index":9,"is_internal_anchor":false},{"citing_arxiv_id":"2605.18365","citing_title":"GeoFlow: Enforcing Implicit Geometric Consistency in Video Generation","ref_index":78,"is_internal_anchor":false},{"citing_arxiv_id":"2507.07982","citing_title":"Geometry Forcing: Marrying Video Diffusion and 3D Representation for Consistent World Modeling","ref_index":79,"is_internal_anchor":false},{"citing_arxiv_id":"2603.11911","citing_title":"InSpatio-WorldFM: An Open-Source Real-Time Generative Frame Model","ref_index":38,"is_internal_anchor":false},{"citing_arxiv_id":"2605.15182","citing_title":"Warp-as-History: Generalizable Camera-Controlled Video Generation from One Training Video","ref_index":10,"is_internal_anchor":false},{"citing_arxiv_id":"2605.09442","citing_title":"SWIFT: Prompt-Adaptive Memory for Efficient Interactive Long Video Generation","ref_index":27,"is_internal_anchor":false},{"citing_arxiv_id":"2604.24764","citing_title":"World-R1: Reinforcing 3D Constraints for Text-to-Video Generation","ref_index":9,"is_internal_anchor":false},{"citing_arxiv_id":"2604.19741","citing_title":"CityRAG: Stepping Into a City via Spatially-Grounded Video Generation","ref_index":64,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/BXELDHNB7GLCBRW66BR2LTWIJA","json":"https://pith.science/pith/BXELDHNB7GLCBRW66BR2LTWIJA.json","graph_json":"https://pith.science/api/pith-number/BXELDHNB7GLCBRW66BR2LTWIJA/graph.json","events_json":"https://pith.science/api/pith-number/BXELDHNB7GLCBRW66BR2LTWIJA/events.json","paper":"https://pith.science/paper/BXELDHNB"},"agent_actions":{"view_html":"https://pith.science/pith/BXELDHNB7GLCBRW66BR2LTWIJA","download_json":"https://pith.science/pith/BXELDHNB7GLCBRW66BR2LTWIJA.json","view_paper":"https://pith.science/paper/BXELDHNB","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2506.05284&json=true","fetch_graph":"https://pith.science/api/pith-number/BXELDHNB7GLCBRW66BR2LTWIJA/graph.json","fetch_events":"https://pith.science/api/pith-number/BXELDHNB7GLCBRW66BR2LTWIJA/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/BXELDHNB7GLCBRW66BR2LTWIJA/action/timestamp_anchor","attest_storage":"https://pith.science/pith/BXELDHNB7GLCBRW66BR2LTWIJA/action/storage_attestation","attest_author":"https://pith.science/pith/BXELDHNB7GLCBRW66BR2LTWIJA/action/author_attestation","sign_citation":"https://pith.science/pith/BXELDHNB7GLCBRW66BR2LTWIJA/action/citation_signature","submit_replication":"https://pith.science/pith/BXELDHNB7GLCBRW66BR2LTWIJA/action/replication_record"}},"created_at":"2026-07-05T11:16:45.169136+00:00","updated_at":"2026-07-05T11:16:45.169136+00:00"}