{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:VLWNTSMLFYEUQEE2MFU7MLUZRU","short_pith_number":"pith:VLWNTSML","schema_version":"1.0","canonical_sha256":"aaecd9c98b2e0948109a6169f62e998d27135dadde20c67356d0786cc984d66e","source":{"kind":"arxiv","id":"2506.05328","version":2},"attestation_state":"computed","paper":{"title":"AV-Reasoner: Improving and Benchmarking Clue-Grounded Audio-Visual Counting for MLLMs","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Guo Chen, Lidong Lu, Tong Lu, Yicheng Liu, Zhiqi Li","submitted_at":"2025-06-05T17:58:33Z","abstract_excerpt":"Despite progress in video understanding, current MLLMs struggle with counting tasks. Existing benchmarks are limited by short videos, close-set queries, lack of clue annotations, and weak multimodal coverage. In this paper, we introduce CG-AV-Counting, a manually-annotated clue-grounded counting benchmark with 1,027 multimodal questions and 5,845 annotated clues over 497 long videos. It supports both black-box and white-box evaluation, serving as a comprehensive testbed for both end-to-end and reasoning-based counting. To explore ways to improve model's counting capability, we propose AV-Reaso"},"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":"2506.05328","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2025-06-05T17:58:33Z","cross_cats_sorted":[],"title_canon_sha256":"a45cde2fc901e345f8a64f3e90786433d14db5a24c4f81782501e2091e8b9c82","abstract_canon_sha256":"a25e182a8eeb7536a45de553f363127f77cab72ea678dab328a3a357c0f68b28"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:40:55.284629Z","signature_b64":"ovT7ffqddHCE3tUzNsmGnx7mvowOdhhIN3IvwMQBjvmPTMFEYLzaC5+0p4bdxyxlwrtVunmcRTX0IYxcvk9eDg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"aaecd9c98b2e0948109a6169f62e998d27135dadde20c67356d0786cc984d66e","last_reissued_at":"2026-07-05T11:40:55.284005Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:40:55.284005Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"AV-Reasoner: Improving and Benchmarking Clue-Grounded Audio-Visual Counting for MLLMs","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Guo Chen, Lidong Lu, Tong Lu, Yicheng Liu, Zhiqi Li","submitted_at":"2025-06-05T17:58:33Z","abstract_excerpt":"Despite progress in video understanding, current MLLMs struggle with counting tasks. Existing benchmarks are limited by short videos, close-set queries, lack of clue annotations, and weak multimodal coverage. In this paper, we introduce CG-AV-Counting, a manually-annotated clue-grounded counting benchmark with 1,027 multimodal questions and 5,845 annotated clues over 497 long videos. It supports both black-box and white-box evaluation, serving as a comprehensive testbed for both end-to-end and reasoning-based counting. To explore ways to improve model's counting capability, we propose AV-Reaso"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2506.05328","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/2506.05328/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":"2506.05328","created_at":"2026-07-05T11:40:55.284079+00:00"},{"alias_kind":"arxiv_version","alias_value":"2506.05328v2","created_at":"2026-07-05T11:40:55.284079+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2506.05328","created_at":"2026-07-05T11:40:55.284079+00:00"},{"alias_kind":"pith_short_12","alias_value":"VLWNTSMLFYEU","created_at":"2026-07-05T11:40:55.284079+00:00"},{"alias_kind":"pith_short_16","alias_value":"VLWNTSMLFYEUQEE2","created_at":"2026-07-05T11:40:55.284079+00:00"},{"alias_kind":"pith_short_8","alias_value":"VLWNTSML","created_at":"2026-07-05T11:40:55.284079+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":5,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2510.15148","citing_title":"XModBench: Benchmarking Cross-Modal Capabilities and Consistency in Omni-Language Models","ref_index":13,"is_internal_anchor":false},{"citing_arxiv_id":"2605.12034","citing_title":"Boosting Omni-Modal Language Models: Staged Post-Training with Visually Debiased Evaluation","ref_index":33,"is_internal_anchor":false},{"citing_arxiv_id":"2605.12034","citing_title":"Boosting Omni-Modal Language Models: Staged Post-Training with Visually Debiased Evaluation","ref_index":33,"is_internal_anchor":false},{"citing_arxiv_id":"2604.14520","citing_title":"Chain of Modality: From Static Fusion to Dynamic Orchestration in Omni-MLLMs","ref_index":24,"is_internal_anchor":false},{"citing_arxiv_id":"2604.16617","citing_title":"AVRT: Audio-Visual Reasoning Transfer through Single-Modality Teachers","ref_index":19,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/VLWNTSMLFYEUQEE2MFU7MLUZRU","json":"https://pith.science/pith/VLWNTSMLFYEUQEE2MFU7MLUZRU.json","graph_json":"https://pith.science/api/pith-number/VLWNTSMLFYEUQEE2MFU7MLUZRU/graph.json","events_json":"https://pith.science/api/pith-number/VLWNTSMLFYEUQEE2MFU7MLUZRU/events.json","paper":"https://pith.science/paper/VLWNTSML"},"agent_actions":{"view_html":"https://pith.science/pith/VLWNTSMLFYEUQEE2MFU7MLUZRU","download_json":"https://pith.science/pith/VLWNTSMLFYEUQEE2MFU7MLUZRU.json","view_paper":"https://pith.science/paper/VLWNTSML","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2506.05328&json=true","fetch_graph":"https://pith.science/api/pith-number/VLWNTSMLFYEUQEE2MFU7MLUZRU/graph.json","fetch_events":"https://pith.science/api/pith-number/VLWNTSMLFYEUQEE2MFU7MLUZRU/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/VLWNTSMLFYEUQEE2MFU7MLUZRU/action/timestamp_anchor","attest_storage":"https://pith.science/pith/VLWNTSMLFYEUQEE2MFU7MLUZRU/action/storage_attestation","attest_author":"https://pith.science/pith/VLWNTSMLFYEUQEE2MFU7MLUZRU/action/author_attestation","sign_citation":"https://pith.science/pith/VLWNTSMLFYEUQEE2MFU7MLUZRU/action/citation_signature","submit_replication":"https://pith.science/pith/VLWNTSMLFYEUQEE2MFU7MLUZRU/action/replication_record"}},"created_at":"2026-07-05T11:40:55.284079+00:00","updated_at":"2026-07-05T11:40:55.284079+00:00"}