{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:XCYCZEAFRJ225HBWZWGMBCERN5","short_pith_number":"pith:XCYCZEAF","schema_version":"1.0","canonical_sha256":"b8b02c90058a75ae9c36cd8cc088916f77f790e8bb1d67b88762a657acc84b44","source":{"kind":"arxiv","id":"2403.16998","version":5},"attestation_state":"computed","paper":{"title":"Understanding Long Videos with Multimodal Language Models","license":"http://creativecommons.org/publicdomain/zero/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Kanchana Ranasinghe, Kumara Kahatapitiya, Michael S. Ryoo, Xiang Li","submitted_at":"2024-03-25T17:59:09Z","abstract_excerpt":"Large Language Models (LLMs) have allowed recent LLM-based approaches to achieve excellent performance on long-video understanding benchmarks. We investigate how extensive world knowledge and strong reasoning skills of underlying LLMs influence this strong performance. Surprisingly, we discover that LLM-based approaches can yield surprisingly good accuracy on long-video tasks with limited video information, sometimes even with no video specific information. Building on this, we explore injecting video-specific information into an LLM-based framework. We utilize off-the-shelf vision tools to ex"},"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":"2403.16998","kind":"arxiv","version":5},"metadata":{"license":"http://creativecommons.org/publicdomain/zero/1.0/","primary_cat":"cs.CV","submitted_at":"2024-03-25T17:59:09Z","cross_cats_sorted":[],"title_canon_sha256":"e9a219a6fb41ca92755a16caff40e8dee298302fbc2ae7f1b661d25c1378b53e","abstract_canon_sha256":"cd3fc319bbd971428dda16d85613b8bf653beb3359fc143cb1edf96b67f8bd66"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:19:48.550795Z","signature_b64":"K4VkEGoZZNkbYOAp4FYJNDg+Zhbc1mk6bLlsLH/lxIZwy+qmSkv1iKupeolXDU6Fk05vAsrk3lVdPr4rQPFHAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"b8b02c90058a75ae9c36cd8cc088916f77f790e8bb1d67b88762a657acc84b44","last_reissued_at":"2026-07-05T11:19:48.550358Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:19:48.550358Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Understanding Long Videos with Multimodal Language Models","license":"http://creativecommons.org/publicdomain/zero/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Kanchana Ranasinghe, Kumara Kahatapitiya, Michael S. Ryoo, Xiang Li","submitted_at":"2024-03-25T17:59:09Z","abstract_excerpt":"Large Language Models (LLMs) have allowed recent LLM-based approaches to achieve excellent performance on long-video understanding benchmarks. We investigate how extensive world knowledge and strong reasoning skills of underlying LLMs influence this strong performance. Surprisingly, we discover that LLM-based approaches can yield surprisingly good accuracy on long-video tasks with limited video information, sometimes even with no video specific information. Building on this, we explore injecting video-specific information into an LLM-based framework. We utilize off-the-shelf vision tools to ex"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2403.16998","kind":"arxiv","version":5},"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/2403.16998/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":"2403.16998","created_at":"2026-07-05T11:19:48.550415+00:00"},{"alias_kind":"arxiv_version","alias_value":"2403.16998v5","created_at":"2026-07-05T11:19:48.550415+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2403.16998","created_at":"2026-07-05T11:19:48.550415+00:00"},{"alias_kind":"pith_short_12","alias_value":"XCYCZEAFRJ22","created_at":"2026-07-05T11:19:48.550415+00:00"},{"alias_kind":"pith_short_16","alias_value":"XCYCZEAFRJ225HBW","created_at":"2026-07-05T11:19:48.550415+00:00"},{"alias_kind":"pith_short_8","alias_value":"XCYCZEAF","created_at":"2026-07-05T11:19:48.550415+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":5,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.23256","citing_title":"P-JEPA: Procedural Video Representation Learning via Joint Embedding Predictive Architecture","ref_index":37,"is_internal_anchor":false},{"citing_arxiv_id":"2604.14692","citing_title":"Chain-of-Glimpse: Search-Guided Progressive Object-Grounded Reasoning for Video Understanding","ref_index":2,"is_internal_anchor":false},{"citing_arxiv_id":"2604.02891","citing_title":"Progressive Video Condensation with MLLM Agent for Long-form Video Understanding","ref_index":26,"is_internal_anchor":false},{"citing_arxiv_id":"2605.08974","citing_title":"Tracking the Truth: Object-Centric Spatio-Temporal Monitoring for Video Large Language Models","ref_index":22,"is_internal_anchor":false},{"citing_arxiv_id":"2604.14692","citing_title":"Chain-of-Glimpse: Search-Guided Progressive Object-Grounded Reasoning for Video Understanding","ref_index":2,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/XCYCZEAFRJ225HBWZWGMBCERN5","json":"https://pith.science/pith/XCYCZEAFRJ225HBWZWGMBCERN5.json","graph_json":"https://pith.science/api/pith-number/XCYCZEAFRJ225HBWZWGMBCERN5/graph.json","events_json":"https://pith.science/api/pith-number/XCYCZEAFRJ225HBWZWGMBCERN5/events.json","paper":"https://pith.science/paper/XCYCZEAF"},"agent_actions":{"view_html":"https://pith.science/pith/XCYCZEAFRJ225HBWZWGMBCERN5","download_json":"https://pith.science/pith/XCYCZEAFRJ225HBWZWGMBCERN5.json","view_paper":"https://pith.science/paper/XCYCZEAF","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2403.16998&json=true","fetch_graph":"https://pith.science/api/pith-number/XCYCZEAFRJ225HBWZWGMBCERN5/graph.json","fetch_events":"https://pith.science/api/pith-number/XCYCZEAFRJ225HBWZWGMBCERN5/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/XCYCZEAFRJ225HBWZWGMBCERN5/action/timestamp_anchor","attest_storage":"https://pith.science/pith/XCYCZEAFRJ225HBWZWGMBCERN5/action/storage_attestation","attest_author":"https://pith.science/pith/XCYCZEAFRJ225HBWZWGMBCERN5/action/author_attestation","sign_citation":"https://pith.science/pith/XCYCZEAFRJ225HBWZWGMBCERN5/action/citation_signature","submit_replication":"https://pith.science/pith/XCYCZEAFRJ225HBWZWGMBCERN5/action/replication_record"}},"created_at":"2026-07-05T11:19:48.550415+00:00","updated_at":"2026-07-05T11:19:48.550415+00:00"}