{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:L7TOXXD3WCCZ2RMCPT72JNSNI2","short_pith_number":"pith:L7TOXXD3","schema_version":"1.0","canonical_sha256":"5fe6ebdc7bb0859d45827cffa4b64d46b1e81cf5d1f62d5e9935490f61fed5c3","source":{"kind":"arxiv","id":"2406.15704","version":1},"attestation_state":"computed","paper":{"title":"video-SALMONN: Speech-Enhanced Audio-Visual Large Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Changli Tang, Chao Zhang, Guangzhi Sun, Lu Lu, Tian Tan, Wei Li, Wenyi Yu, Xianzhao Chen, Yuxuan Wang, Zejun Ma","submitted_at":"2024-06-22T01:36:11Z","abstract_excerpt":"Speech understanding as an element of the more generic video understanding using audio-visual large language models (av-LLMs) is a crucial yet understudied aspect. This paper proposes video-SALMONN, a single end-to-end av-LLM for video processing, which can understand not only visual frame sequences, audio events and music, but speech as well. To obtain fine-grained temporal information required by speech understanding, while keeping efficient for other video elements, this paper proposes a novel multi-resolution causal Q-Former (MRC Q-Former) structure to connect pre-trained audio-visual enco"},"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":"2406.15704","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2024-06-22T01:36:11Z","cross_cats_sorted":[],"title_canon_sha256":"f9852db2f8430da1c8763ad4d9af924f9a75e2d45901f47119e4e93b9ef02ce4","abstract_canon_sha256":"bfe3e1126f7d0a218f87b6bad67b10fbfc746b0c10f8dd5ef445adbc71404f5d"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:35:26.326906Z","signature_b64":"7r4MWhTHry28HKN9iTTqZ3+57xBg0U+e8ens9a6r1bml8Dua1W8bYMoxL9oW4V2YzYQXA7a7K906fKtGIUTmAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"5fe6ebdc7bb0859d45827cffa4b64d46b1e81cf5d1f62d5e9935490f61fed5c3","last_reissued_at":"2026-07-05T08:35:26.326395Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:35:26.326395Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"video-SALMONN: Speech-Enhanced Audio-Visual Large Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Changli Tang, Chao Zhang, Guangzhi Sun, Lu Lu, Tian Tan, Wei Li, Wenyi Yu, Xianzhao Chen, Yuxuan Wang, Zejun Ma","submitted_at":"2024-06-22T01:36:11Z","abstract_excerpt":"Speech understanding as an element of the more generic video understanding using audio-visual large language models (av-LLMs) is a crucial yet understudied aspect. This paper proposes video-SALMONN, a single end-to-end av-LLM for video processing, which can understand not only visual frame sequences, audio events and music, but speech as well. To obtain fine-grained temporal information required by speech understanding, while keeping efficient for other video elements, this paper proposes a novel multi-resolution causal Q-Former (MRC Q-Former) structure to connect pre-trained audio-visual enco"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2406.15704","kind":"arxiv","version":1},"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/2406.15704/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":"2406.15704","created_at":"2026-07-05T08:35:26.326457+00:00"},{"alias_kind":"arxiv_version","alias_value":"2406.15704v1","created_at":"2026-07-05T08:35:26.326457+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2406.15704","created_at":"2026-07-05T08:35:26.326457+00:00"},{"alias_kind":"pith_short_12","alias_value":"L7TOXXD3WCCZ","created_at":"2026-07-05T08:35:26.326457+00:00"},{"alias_kind":"pith_short_16","alias_value":"L7TOXXD3WCCZ2RMC","created_at":"2026-07-05T08:35:26.326457+00:00"},{"alias_kind":"pith_short_8","alias_value":"L7TOXXD3","created_at":"2026-07-05T08:35:26.326457+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":15,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.21949","citing_title":"CapRiCorn-1K: A Comprehensive Benchmark for Video Captioning and Subject Referential Consistency Across Temporal Scales","ref_index":9,"is_internal_anchor":false},{"citing_arxiv_id":"2606.00579","citing_title":"Sandboxed Coding Agents are Competitive Omni-modal Task Solvers","ref_index":6,"is_internal_anchor":false},{"citing_arxiv_id":"2606.00508","citing_title":"V-LynX: Token Interface Alignment for Video+X LLMs","ref_index":7,"is_internal_anchor":false},{"citing_arxiv_id":"2510.15148","citing_title":"XModBench: Benchmarking Cross-Modal Capabilities and Consistency in Omni-Language Models","ref_index":17,"is_internal_anchor":false},{"citing_arxiv_id":"2501.01957","citing_title":"VITA-1.5: Towards GPT-4o Level Real-Time Vision and Speech Interaction","ref_index":5,"is_internal_anchor":false},{"citing_arxiv_id":"2511.14582","citing_title":"OmniZip: Audio-Guided Dynamic Token Compression for Fast Omnimodal Large Language Models","ref_index":38,"is_internal_anchor":false},{"citing_arxiv_id":"2502.04326","citing_title":"WorldSense: Evaluating Real-world Omnimodal Understanding for Multimodal LLMs","ref_index":58,"is_internal_anchor":false},{"citing_arxiv_id":"2512.02231","citing_title":"See, Hear, and Understand: Benchmarking Audiovisual Human Speech Understanding in Multimodal Large Language Models","ref_index":52,"is_internal_anchor":false},{"citing_arxiv_id":"2503.12937","citing_title":"R1-VL: Learning to Reason with Multimodal Large Language Models via Step-wise Group Relative Policy Optimization","ref_index":35,"is_internal_anchor":false},{"citing_arxiv_id":"2604.02605","citing_title":"Do Audio-Visual Large Language Models Really See and Hear?","ref_index":51,"is_internal_anchor":false},{"citing_arxiv_id":"2605.12056","citing_title":"OmniRefine: Alignment-Aware Cooperative Compression for Efficient Omnimodal Large Language Models","ref_index":42,"is_internal_anchor":false},{"citing_arxiv_id":"2605.01024","citing_title":"EmoMM: Benchmarking and Steering MLLM for Multimodal Emotion Recognition under Conflict and Missingness","ref_index":74,"is_internal_anchor":false},{"citing_arxiv_id":"2604.11244","citing_title":"Script-a-Video: Deep Structured Audio-visual Captions via Factorized Streams and Relational Grounding","ref_index":21,"is_internal_anchor":false},{"citing_arxiv_id":"2501.13106","citing_title":"VideoLLaMA 3: Frontier Multimodal Foundation Models for Image and Video Understanding","ref_index":24,"is_internal_anchor":false},{"citing_arxiv_id":"2503.20215","citing_title":"Qwen2.5-Omni Technical Report","ref_index":33,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/L7TOXXD3WCCZ2RMCPT72JNSNI2","json":"https://pith.science/pith/L7TOXXD3WCCZ2RMCPT72JNSNI2.json","graph_json":"https://pith.science/api/pith-number/L7TOXXD3WCCZ2RMCPT72JNSNI2/graph.json","events_json":"https://pith.science/api/pith-number/L7TOXXD3WCCZ2RMCPT72JNSNI2/events.json","paper":"https://pith.science/paper/L7TOXXD3"},"agent_actions":{"view_html":"https://pith.science/pith/L7TOXXD3WCCZ2RMCPT72JNSNI2","download_json":"https://pith.science/pith/L7TOXXD3WCCZ2RMCPT72JNSNI2.json","view_paper":"https://pith.science/paper/L7TOXXD3","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2406.15704&json=true","fetch_graph":"https://pith.science/api/pith-number/L7TOXXD3WCCZ2RMCPT72JNSNI2/graph.json","fetch_events":"https://pith.science/api/pith-number/L7TOXXD3WCCZ2RMCPT72JNSNI2/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/L7TOXXD3WCCZ2RMCPT72JNSNI2/action/timestamp_anchor","attest_storage":"https://pith.science/pith/L7TOXXD3WCCZ2RMCPT72JNSNI2/action/storage_attestation","attest_author":"https://pith.science/pith/L7TOXXD3WCCZ2RMCPT72JNSNI2/action/author_attestation","sign_citation":"https://pith.science/pith/L7TOXXD3WCCZ2RMCPT72JNSNI2/action/citation_signature","submit_replication":"https://pith.science/pith/L7TOXXD3WCCZ2RMCPT72JNSNI2/action/replication_record"}},"created_at":"2026-07-05T08:35:26.326457+00:00","updated_at":"2026-07-05T08:35:26.326457+00:00"}