{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:OHG6RQEA4BFIPGX2N5VSLOJIK2","short_pith_number":"pith:OHG6RQEA","schema_version":"1.0","canonical_sha256":"71cde8c080e04a879afa6f6b25b928569850845d1852d92528c3023510a5b25d","source":{"kind":"arxiv","id":"2502.12632","version":3},"attestation_state":"computed","paper":{"title":"MALT Diffusion: Memory-Augmented Latent Transformers for Any-Length Video Generation","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CV","authors_text":"Agrim Gupta, Dan Kondratyuk, David Ross, Irfan Essa, Jinwoo Shin, Jonathan Huang, Jos\\'e Lezama, Meera Hahn, Sihyun Yu","submitted_at":"2025-02-18T08:22:50Z","abstract_excerpt":"Diffusion models are successful for synthesizing high-quality videos but are limited to generating short clips (e.g., 2-10 seconds). Synthesizing sustained footage (e.g. over minutes) still remains an open research question. In this paper, we propose MALT Diffusion (using Memory-Augmented Latent Transformers), a new diffusion model specialized for long video generation. MALT Diffusion (or just MALT) handles long videos by subdividing them into short segments and doing segment-level autoregressive generation. To achieve this, we first propose recurrent attention layers that encode multiple segm"},"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":"2502.12632","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2025-02-18T08:22:50Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"5efe78921e0aed4e37ce0530e0dc3386c510c7f4da88753de7b0f823fdb2428d","abstract_canon_sha256":"c2fc3177c77d08a5697ffa23e67a89bb3ba243f379395eeea033ff554127909b"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:33:20.134420Z","signature_b64":"NU5AqUgljATGWpjmOTA8yl3DpMeFNUHGUZrUc8itAr/fwaScccJNhTtLJsw1fBI6792BBkgXgT0sDNqZU1BvDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"71cde8c080e04a879afa6f6b25b928569850845d1852d92528c3023510a5b25d","last_reissued_at":"2026-07-05T11:33:20.133851Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:33:20.133851Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"MALT Diffusion: Memory-Augmented Latent Transformers for Any-Length Video Generation","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CV","authors_text":"Agrim Gupta, Dan Kondratyuk, David Ross, Irfan Essa, Jinwoo Shin, Jonathan Huang, Jos\\'e Lezama, Meera Hahn, Sihyun Yu","submitted_at":"2025-02-18T08:22:50Z","abstract_excerpt":"Diffusion models are successful for synthesizing high-quality videos but are limited to generating short clips (e.g., 2-10 seconds). Synthesizing sustained footage (e.g. over minutes) still remains an open research question. In this paper, we propose MALT Diffusion (using Memory-Augmented Latent Transformers), a new diffusion model specialized for long video generation. MALT Diffusion (or just MALT) handles long videos by subdividing them into short segments and doing segment-level autoregressive generation. To achieve this, we first propose recurrent attention layers that encode multiple segm"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2502.12632","kind":"arxiv","version":3},"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/2502.12632/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":"2502.12632","created_at":"2026-07-05T11:33:20.133910+00:00"},{"alias_kind":"arxiv_version","alias_value":"2502.12632v3","created_at":"2026-07-05T11:33:20.133910+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2502.12632","created_at":"2026-07-05T11:33:20.133910+00:00"},{"alias_kind":"pith_short_12","alias_value":"OHG6RQEA4BFI","created_at":"2026-07-05T11:33:20.133910+00:00"},{"alias_kind":"pith_short_16","alias_value":"OHG6RQEA4BFIPGX2","created_at":"2026-07-05T11:33:20.133910+00:00"},{"alias_kind":"pith_short_8","alias_value":"OHG6RQEA","created_at":"2026-07-05T11:33:20.133910+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.17590","citing_title":"TivTok: Broadcasting Time-Invariant Tokens for Scalable Video Tokenization","ref_index":125,"is_internal_anchor":false},{"citing_arxiv_id":"2605.15042","citing_title":"EverAnimate: Minute-Scale Human Animation via Latent Flow Restoration","ref_index":53,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/OHG6RQEA4BFIPGX2N5VSLOJIK2","json":"https://pith.science/pith/OHG6RQEA4BFIPGX2N5VSLOJIK2.json","graph_json":"https://pith.science/api/pith-number/OHG6RQEA4BFIPGX2N5VSLOJIK2/graph.json","events_json":"https://pith.science/api/pith-number/OHG6RQEA4BFIPGX2N5VSLOJIK2/events.json","paper":"https://pith.science/paper/OHG6RQEA"},"agent_actions":{"view_html":"https://pith.science/pith/OHG6RQEA4BFIPGX2N5VSLOJIK2","download_json":"https://pith.science/pith/OHG6RQEA4BFIPGX2N5VSLOJIK2.json","view_paper":"https://pith.science/paper/OHG6RQEA","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2502.12632&json=true","fetch_graph":"https://pith.science/api/pith-number/OHG6RQEA4BFIPGX2N5VSLOJIK2/graph.json","fetch_events":"https://pith.science/api/pith-number/OHG6RQEA4BFIPGX2N5VSLOJIK2/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/OHG6RQEA4BFIPGX2N5VSLOJIK2/action/timestamp_anchor","attest_storage":"https://pith.science/pith/OHG6RQEA4BFIPGX2N5VSLOJIK2/action/storage_attestation","attest_author":"https://pith.science/pith/OHG6RQEA4BFIPGX2N5VSLOJIK2/action/author_attestation","sign_citation":"https://pith.science/pith/OHG6RQEA4BFIPGX2N5VSLOJIK2/action/citation_signature","submit_replication":"https://pith.science/pith/OHG6RQEA4BFIPGX2N5VSLOJIK2/action/replication_record"}},"created_at":"2026-07-05T11:33:20.133910+00:00","updated_at":"2026-07-05T11:33:20.133910+00:00"}