{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:67CMYWP3I7ICIBO6FNPHV64WZD","short_pith_number":"pith:67CMYWP3","schema_version":"1.0","canonical_sha256":"f7c4cc59fb47d02405de2b5e7afb96c8c54cbefccfdf92ee331a32e209f2e10f","source":{"kind":"arxiv","id":"2512.05774","version":2},"attestation_state":"computed","paper":{"title":"Active Video Perception: Iterative Evidence Seeking for Agentic Long Video Understanding","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.CV","authors_text":"Caiming Xiong, Honglu Zhou, Juan Carlos Niebles, Junnan Li, Michael S. Ryoo, Mohit Bansal, Shijie Wang, Silvio Savarese, Ziyang Wang","submitted_at":"2025-12-05T15:03:48Z","abstract_excerpt":"Long video understanding (LVU) is challenging because answering real-world queries often depends on sparse, temporally dispersed cues buried in hours of mostly redundant and irrelevant content. While agentic pipelines improve video reasoning capabilities, prevailing frameworks rely on a query-agnostic captioner to perceive video information, which wastes computation on irrelevant content and blurs fine-grained temporal and spatial information. Motivated by active perception theory, we argue that LVU agents should actively decide what, when, and where to observe, and continuously assess whether"},"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":"2512.05774","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.CV","submitted_at":"2025-12-05T15:03:48Z","cross_cats_sorted":["cs.AI","cs.CL"],"title_canon_sha256":"7c2b8020785f02fcf11a78bae811df2f0bbd4773bc90b83cde973b0211c5afc3","abstract_canon_sha256":"cb01c5a89119f1f05fbc464352bf9fe04fb170b66efe64baf06622139fcddce8"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-05T01:14:31.807105Z","signature_b64":"shmCMTScL8xUTh4evacQFpuUKwhfnzKluD4ENkSKg/gxrPspjYqn6ZZcrZ+zlWiZxbCqp3+jkDC70mFErPeVCQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"f7c4cc59fb47d02405de2b5e7afb96c8c54cbefccfdf92ee331a32e209f2e10f","last_reissued_at":"2026-06-05T01:14:31.806447Z","signature_status":"signed_v1","first_computed_at":"2026-06-05T01:14:31.806447Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Active Video Perception: Iterative Evidence Seeking for Agentic Long Video Understanding","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.CV","authors_text":"Caiming Xiong, Honglu Zhou, Juan Carlos Niebles, Junnan Li, Michael S. Ryoo, Mohit Bansal, Shijie Wang, Silvio Savarese, Ziyang Wang","submitted_at":"2025-12-05T15:03:48Z","abstract_excerpt":"Long video understanding (LVU) is challenging because answering real-world queries often depends on sparse, temporally dispersed cues buried in hours of mostly redundant and irrelevant content. While agentic pipelines improve video reasoning capabilities, prevailing frameworks rely on a query-agnostic captioner to perceive video information, which wastes computation on irrelevant content and blurs fine-grained temporal and spatial information. Motivated by active perception theory, we argue that LVU agents should actively decide what, when, and where to observe, and continuously assess whether"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2512.05774","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/2512.05774/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":"2512.05774","created_at":"2026-06-05T01:14:31.806527+00:00"},{"alias_kind":"arxiv_version","alias_value":"2512.05774v2","created_at":"2026-06-05T01:14:31.806527+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2512.05774","created_at":"2026-06-05T01:14:31.806527+00:00"},{"alias_kind":"pith_short_12","alias_value":"67CMYWP3I7IC","created_at":"2026-06-05T01:14:31.806527+00:00"},{"alias_kind":"pith_short_16","alias_value":"67CMYWP3I7ICIBO6","created_at":"2026-06-05T01:14:31.806527+00:00"},{"alias_kind":"pith_short_8","alias_value":"67CMYWP3","created_at":"2026-06-05T01:14:31.806527+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":3,"internal_anchor_count":3,"sample":[{"citing_arxiv_id":"2605.23826","citing_title":"Decomposing Queries into Tool Calls for Long-Video Keyframe Retrieval","ref_index":21,"is_internal_anchor":true},{"citing_arxiv_id":"2605.10936","citing_title":"Personal Visual Context Learning in Large Multimodal Models","ref_index":75,"is_internal_anchor":true},{"citing_arxiv_id":"2605.09874","citing_title":"EgoMemReason: A Memory-Driven Reasoning Benchmark for Long-Horizon Egocentric Video Understanding","ref_index":72,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/67CMYWP3I7ICIBO6FNPHV64WZD","json":"https://pith.science/pith/67CMYWP3I7ICIBO6FNPHV64WZD.json","graph_json":"https://pith.science/api/pith-number/67CMYWP3I7ICIBO6FNPHV64WZD/graph.json","events_json":"https://pith.science/api/pith-number/67CMYWP3I7ICIBO6FNPHV64WZD/events.json","paper":"https://pith.science/paper/67CMYWP3"},"agent_actions":{"view_html":"https://pith.science/pith/67CMYWP3I7ICIBO6FNPHV64WZD","download_json":"https://pith.science/pith/67CMYWP3I7ICIBO6FNPHV64WZD.json","view_paper":"https://pith.science/paper/67CMYWP3","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2512.05774&json=true","fetch_graph":"https://pith.science/api/pith-number/67CMYWP3I7ICIBO6FNPHV64WZD/graph.json","fetch_events":"https://pith.science/api/pith-number/67CMYWP3I7ICIBO6FNPHV64WZD/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/67CMYWP3I7ICIBO6FNPHV64WZD/action/timestamp_anchor","attest_storage":"https://pith.science/pith/67CMYWP3I7ICIBO6FNPHV64WZD/action/storage_attestation","attest_author":"https://pith.science/pith/67CMYWP3I7ICIBO6FNPHV64WZD/action/author_attestation","sign_citation":"https://pith.science/pith/67CMYWP3I7ICIBO6FNPHV64WZD/action/citation_signature","submit_replication":"https://pith.science/pith/67CMYWP3I7ICIBO6FNPHV64WZD/action/replication_record"}},"created_at":"2026-06-05T01:14:31.806527+00:00","updated_at":"2026-06-05T01:14:31.806527+00:00"}