{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:N3EL6NTDX4KD6KF432R4IUMC5X","short_pith_number":"pith:N3EL6NTD","schema_version":"1.0","canonical_sha256":"6ec8bf3663bf143f28bcdea3c45182edcd90b53d38e4a6bea0388b22f9ba64b9","source":{"kind":"arxiv","id":"2402.05861","version":2},"attestation_state":"computed","paper":{"title":"Memory Consolidation Enables Long-Context Video Understanding","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Ivana Bala\\v{z}evi\\'c, Olivier J. H\\'enaff, Pinelopi Papalampidi, Rahma Chaabouni, Skanda Koppula, Yuge Shi","submitted_at":"2024-02-08T17:50:22Z","abstract_excerpt":"Most transformer-based video encoders are limited to short temporal contexts due to their quadratic complexity. While various attempts have been made to extend this context, this has often come at the cost of both conceptual and computational complexity. We propose to instead re-purpose existing pre-trained video transformers by simply fine-tuning them to attend to memories derived non-parametrically from past activations. By leveraging redundancy reduction, our memory-consolidated vision transformer (MC-ViT) effortlessly extends its context far into the past and exhibits excellent scaling beh"},"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":"2402.05861","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2024-02-08T17:50:22Z","cross_cats_sorted":[],"title_canon_sha256":"ed59edf4439dad57428dc85bd843542989e5fcfa15a5c2fa8cabf18e88410b88","abstract_canon_sha256":"3adb473aec7b19705d10d4c5761fe464de024d272e3cb0f87e8d9b5ad4ef5d55"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:25:31.602132Z","signature_b64":"HuIQQ7nq83XjSq2uhwA3nBcrwFLHy/pR/unnSTAHd+gGItH/hEFtmZC6oMxEOrae2WtEL4eOEXBM3Gi3AiRsCg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"6ec8bf3663bf143f28bcdea3c45182edcd90b53d38e4a6bea0388b22f9ba64b9","last_reissued_at":"2026-07-05T08:25:31.601668Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:25:31.601668Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Memory Consolidation Enables Long-Context Video Understanding","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Ivana Bala\\v{z}evi\\'c, Olivier J. H\\'enaff, Pinelopi Papalampidi, Rahma Chaabouni, Skanda Koppula, Yuge Shi","submitted_at":"2024-02-08T17:50:22Z","abstract_excerpt":"Most transformer-based video encoders are limited to short temporal contexts due to their quadratic complexity. While various attempts have been made to extend this context, this has often come at the cost of both conceptual and computational complexity. We propose to instead re-purpose existing pre-trained video transformers by simply fine-tuning them to attend to memories derived non-parametrically from past activations. By leveraging redundancy reduction, our memory-consolidated vision transformer (MC-ViT) effortlessly extends its context far into the past and exhibits excellent scaling beh"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2402.05861","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/2402.05861/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":"2402.05861","created_at":"2026-07-05T08:25:31.601726+00:00"},{"alias_kind":"arxiv_version","alias_value":"2402.05861v2","created_at":"2026-07-05T08:25:31.601726+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2402.05861","created_at":"2026-07-05T08:25:31.601726+00:00"},{"alias_kind":"pith_short_12","alias_value":"N3EL6NTDX4KD","created_at":"2026-07-05T08:25:31.601726+00:00"},{"alias_kind":"pith_short_16","alias_value":"N3EL6NTDX4KD6KF4","created_at":"2026-07-05T08:25:31.601726+00:00"},{"alias_kind":"pith_short_8","alias_value":"N3EL6NTD","created_at":"2026-07-05T08:25:31.601726+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":3,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2607.08032","citing_title":"What to Keep, What to Forget: A Rate--Distortion View of Memory Compaction in LLMs and Agents","ref_index":4,"is_internal_anchor":true},{"citing_arxiv_id":"2605.00814","citing_title":"Persistent Visual Memory: Sustaining Perception for Deep Generation in LVLMs","ref_index":7,"is_internal_anchor":false},{"citing_arxiv_id":"2605.00814","citing_title":"Persistent Visual Memory: Sustaining Perception for Deep Generation in LVLMs","ref_index":7,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/N3EL6NTDX4KD6KF432R4IUMC5X","json":"https://pith.science/pith/N3EL6NTDX4KD6KF432R4IUMC5X.json","graph_json":"https://pith.science/api/pith-number/N3EL6NTDX4KD6KF432R4IUMC5X/graph.json","events_json":"https://pith.science/api/pith-number/N3EL6NTDX4KD6KF432R4IUMC5X/events.json","paper":"https://pith.science/paper/N3EL6NTD"},"agent_actions":{"view_html":"https://pith.science/pith/N3EL6NTDX4KD6KF432R4IUMC5X","download_json":"https://pith.science/pith/N3EL6NTDX4KD6KF432R4IUMC5X.json","view_paper":"https://pith.science/paper/N3EL6NTD","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2402.05861&json=true","fetch_graph":"https://pith.science/api/pith-number/N3EL6NTDX4KD6KF432R4IUMC5X/graph.json","fetch_events":"https://pith.science/api/pith-number/N3EL6NTDX4KD6KF432R4IUMC5X/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/N3EL6NTDX4KD6KF432R4IUMC5X/action/timestamp_anchor","attest_storage":"https://pith.science/pith/N3EL6NTDX4KD6KF432R4IUMC5X/action/storage_attestation","attest_author":"https://pith.science/pith/N3EL6NTDX4KD6KF432R4IUMC5X/action/author_attestation","sign_citation":"https://pith.science/pith/N3EL6NTDX4KD6KF432R4IUMC5X/action/citation_signature","submit_replication":"https://pith.science/pith/N3EL6NTDX4KD6KF432R4IUMC5X/action/replication_record"}},"created_at":"2026-07-05T08:25:31.601726+00:00","updated_at":"2026-07-05T08:25:31.601726+00:00"}