{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:OT53LJ3ZNWR3JVDN2MBRB4OGKA","short_pith_number":"pith:OT53LJ3Z","schema_version":"1.0","canonical_sha256":"74fbb5a7796da3b4d46dd30310f1c6503ad650075c02be55c89c34e2cc56caf9","source":{"kind":"arxiv","id":"2510.01009","version":3},"attestation_state":"computed","paper":{"title":"POVQA: Preference-Optimized Video Question Answering with Rationales for Data Efficiency","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.MM"],"primary_cat":"cs.CV","authors_text":"Ankit Ghimire, Ashim Dahal, Nick Rahimi, Saydul Akbar Murad","submitted_at":"2025-10-01T15:15:36Z","abstract_excerpt":"Long-video multimodal question answering requires structured reasoning over visual evidence and dialogue, but Large Vision-Language Models (LVLMs) are constrained by context-window and compute limits. We propose POVQA, which compresses each second into a temporally pooled image (1 fps pooled images) to maintain dense temporal coverage under a fixed token budget. We then train Qwen2.5-VL-7B with supervised fine-tuning (SFT) on rationale+answer targets, and optionally apply Direct Preference Optimization (DPO) for preference alignment. We introduce ReasonVQA as a pilot diagnostic dataset with 12"},"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":"2510.01009","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2025-10-01T15:15:36Z","cross_cats_sorted":["cs.MM"],"title_canon_sha256":"7a9957a722a8a395be1f401e7cbfe7984072f79c5bddf7b3fc72fe1131f5a6b3","abstract_canon_sha256":"bf693a29bf15a1466c7034658d81b5454456113b024cbddf7aef2548b7767f74"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-02T02:04:09.506138Z","signature_b64":"AfKbA02uUPLGJ8L7d4LheU+3Tk0a9+dGrDz/RVuxzUnF2U09E1rd4sut4M28YK78IZHDhbXUTPc2Kay9F1NbCg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"74fbb5a7796da3b4d46dd30310f1c6503ad650075c02be55c89c34e2cc56caf9","last_reissued_at":"2026-06-02T02:04:09.505625Z","signature_status":"signed_v1","first_computed_at":"2026-06-02T02:04:09.505625Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"POVQA: Preference-Optimized Video Question Answering with Rationales for Data Efficiency","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.MM"],"primary_cat":"cs.CV","authors_text":"Ankit Ghimire, Ashim Dahal, Nick Rahimi, Saydul Akbar Murad","submitted_at":"2025-10-01T15:15:36Z","abstract_excerpt":"Long-video multimodal question answering requires structured reasoning over visual evidence and dialogue, but Large Vision-Language Models (LVLMs) are constrained by context-window and compute limits. We propose POVQA, which compresses each second into a temporally pooled image (1 fps pooled images) to maintain dense temporal coverage under a fixed token budget. We then train Qwen2.5-VL-7B with supervised fine-tuning (SFT) on rationale+answer targets, and optionally apply Direct Preference Optimization (DPO) for preference alignment. We introduce ReasonVQA as a pilot diagnostic dataset with 12"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2510.01009","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/2510.01009/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":"2510.01009","created_at":"2026-06-02T02:04:09.505687+00:00"},{"alias_kind":"arxiv_version","alias_value":"2510.01009v3","created_at":"2026-06-02T02:04:09.505687+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2510.01009","created_at":"2026-06-02T02:04:09.505687+00:00"},{"alias_kind":"pith_short_12","alias_value":"OT53LJ3ZNWR3","created_at":"2026-06-02T02:04:09.505687+00:00"},{"alias_kind":"pith_short_16","alias_value":"OT53LJ3ZNWR3JVDN","created_at":"2026-06-02T02:04:09.505687+00:00"},{"alias_kind":"pith_short_8","alias_value":"OT53LJ3Z","created_at":"2026-06-02T02:04:09.505687+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/OT53LJ3ZNWR3JVDN2MBRB4OGKA","json":"https://pith.science/pith/OT53LJ3ZNWR3JVDN2MBRB4OGKA.json","graph_json":"https://pith.science/api/pith-number/OT53LJ3ZNWR3JVDN2MBRB4OGKA/graph.json","events_json":"https://pith.science/api/pith-number/OT53LJ3ZNWR3JVDN2MBRB4OGKA/events.json","paper":"https://pith.science/paper/OT53LJ3Z"},"agent_actions":{"view_html":"https://pith.science/pith/OT53LJ3ZNWR3JVDN2MBRB4OGKA","download_json":"https://pith.science/pith/OT53LJ3ZNWR3JVDN2MBRB4OGKA.json","view_paper":"https://pith.science/paper/OT53LJ3Z","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2510.01009&json=true","fetch_graph":"https://pith.science/api/pith-number/OT53LJ3ZNWR3JVDN2MBRB4OGKA/graph.json","fetch_events":"https://pith.science/api/pith-number/OT53LJ3ZNWR3JVDN2MBRB4OGKA/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/OT53LJ3ZNWR3JVDN2MBRB4OGKA/action/timestamp_anchor","attest_storage":"https://pith.science/pith/OT53LJ3ZNWR3JVDN2MBRB4OGKA/action/storage_attestation","attest_author":"https://pith.science/pith/OT53LJ3ZNWR3JVDN2MBRB4OGKA/action/author_attestation","sign_citation":"https://pith.science/pith/OT53LJ3ZNWR3JVDN2MBRB4OGKA/action/citation_signature","submit_replication":"https://pith.science/pith/OT53LJ3ZNWR3JVDN2MBRB4OGKA/action/replication_record"}},"created_at":"2026-06-02T02:04:09.505687+00:00","updated_at":"2026-06-02T02:04:09.505687+00:00"}