{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:DNTULYCQAPZB6LAMJJTE2TH74L","short_pith_number":"pith:DNTULYCQ","schema_version":"1.0","canonical_sha256":"1b6745e05003f21f2c0c4a664d4cffe2d9a2eb102458ff7048d7ce4510b77fa5","source":{"kind":"arxiv","id":"2503.08585","version":2},"attestation_state":"computed","paper":{"title":"HierarQ: Task-Aware Hierarchical Q-Former for Enhanced Video Understanding","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Shehreen Azad, Vibhav Vineet, Yogesh Singh Rawat","submitted_at":"2025-03-11T16:21:23Z","abstract_excerpt":"Despite advancements in multimodal large language models (MLLMs), current approaches struggle in medium-to-long video understanding due to frame and context length limitations. As a result, these models often depend on frame sampling, which risks missing key information over time and lacks task-specific relevance. To address these challenges, we introduce HierarQ, a task-aware hierarchical Q-Former based framework that sequentially processes frames to bypass the need for frame sampling, while avoiding LLM's context length limitations. We introduce a lightweight two-stream language-guided featu"},"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":"2503.08585","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2025-03-11T16:21:23Z","cross_cats_sorted":[],"title_canon_sha256":"16c56d651f39fd6025fc16fd3e0dc46daa2620c66583b4a0d0c28fcca09075e2","abstract_canon_sha256":"d56e9e6c21b511d548af8bd38bbcec82c831133d6a23f0296d40120398aded09"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:53:39.422772Z","signature_b64":"dnSgifDQdp3IVMf0vJ9rmyo+3aKXsGiOpm697+cU2UxD11XilQSbXxLiDCvJYE89j36FUtvOvC5vecN7Xk92Cw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"1b6745e05003f21f2c0c4a664d4cffe2d9a2eb102458ff7048d7ce4510b77fa5","last_reissued_at":"2026-07-05T10:53:39.422300Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:53:39.422300Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"HierarQ: Task-Aware Hierarchical Q-Former for Enhanced Video Understanding","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Shehreen Azad, Vibhav Vineet, Yogesh Singh Rawat","submitted_at":"2025-03-11T16:21:23Z","abstract_excerpt":"Despite advancements in multimodal large language models (MLLMs), current approaches struggle in medium-to-long video understanding due to frame and context length limitations. As a result, these models often depend on frame sampling, which risks missing key information over time and lacks task-specific relevance. To address these challenges, we introduce HierarQ, a task-aware hierarchical Q-Former based framework that sequentially processes frames to bypass the need for frame sampling, while avoiding LLM's context length limitations. We introduce a lightweight two-stream language-guided featu"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2503.08585","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/2503.08585/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":"2503.08585","created_at":"2026-07-05T10:53:39.422358+00:00"},{"alias_kind":"arxiv_version","alias_value":"2503.08585v2","created_at":"2026-07-05T10:53:39.422358+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2503.08585","created_at":"2026-07-05T10:53:39.422358+00:00"},{"alias_kind":"pith_short_12","alias_value":"DNTULYCQAPZB","created_at":"2026-07-05T10:53:39.422358+00:00"},{"alias_kind":"pith_short_16","alias_value":"DNTULYCQAPZB6LAM","created_at":"2026-07-05T10:53:39.422358+00:00"},{"alias_kind":"pith_short_8","alias_value":"DNTULYCQ","created_at":"2026-07-05T10:53:39.422358+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2508.10922","citing_title":"A Survey on Video Temporal Grounding with Multimodal Large Language Model","ref_index":70,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/DNTULYCQAPZB6LAMJJTE2TH74L","json":"https://pith.science/pith/DNTULYCQAPZB6LAMJJTE2TH74L.json","graph_json":"https://pith.science/api/pith-number/DNTULYCQAPZB6LAMJJTE2TH74L/graph.json","events_json":"https://pith.science/api/pith-number/DNTULYCQAPZB6LAMJJTE2TH74L/events.json","paper":"https://pith.science/paper/DNTULYCQ"},"agent_actions":{"view_html":"https://pith.science/pith/DNTULYCQAPZB6LAMJJTE2TH74L","download_json":"https://pith.science/pith/DNTULYCQAPZB6LAMJJTE2TH74L.json","view_paper":"https://pith.science/paper/DNTULYCQ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2503.08585&json=true","fetch_graph":"https://pith.science/api/pith-number/DNTULYCQAPZB6LAMJJTE2TH74L/graph.json","fetch_events":"https://pith.science/api/pith-number/DNTULYCQAPZB6LAMJJTE2TH74L/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/DNTULYCQAPZB6LAMJJTE2TH74L/action/timestamp_anchor","attest_storage":"https://pith.science/pith/DNTULYCQAPZB6LAMJJTE2TH74L/action/storage_attestation","attest_author":"https://pith.science/pith/DNTULYCQAPZB6LAMJJTE2TH74L/action/author_attestation","sign_citation":"https://pith.science/pith/DNTULYCQAPZB6LAMJJTE2TH74L/action/citation_signature","submit_replication":"https://pith.science/pith/DNTULYCQAPZB6LAMJJTE2TH74L/action/replication_record"}},"created_at":"2026-07-05T10:53:39.422358+00:00","updated_at":"2026-07-05T10:53:39.422358+00:00"}