{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:TPLPIKCZSSCXR4IOQ3DXDE5EBA","short_pith_number":"pith:TPLPIKCZ","schema_version":"1.0","canonical_sha256":"9bd6f42859948578f10e86c77193a4080f85796e4ae22c26a20d00ef185b30ec","source":{"kind":"arxiv","id":"2605.18041","version":1},"attestation_state":"computed","paper":{"title":"OmniSelect: Dynamic Modality-Aware Token Compression for Efficient Omni-modal Large Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Jianxin Zhang, Juntao Li, Le Li, Morunliu Yang, Peifeng Li, Ruotao Xu, Siwei Feng, Yihang Lou, Yue Wang","submitted_at":"2026-05-18T08:33:52Z","abstract_excerpt":"Omnimodal large language models (OmniLLMs) have recently gained increasing attention for unified audio-video understanding. However, processing long multimodal token sequences introduces substantial computational overhead, making efficient token compression crucial. Existing methods typically rely on fixed, modality-specific guidance, which fails to account for the varying importance of modalities across different queries. To address this limitation, we propose $\\textbf{OmniSelect}$, a training-free, modality-adaptive token pruning framework that dynamically selects appropriate compression str"},"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":"2605.18041","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2026-05-18T08:33:52Z","cross_cats_sorted":[],"title_canon_sha256":"4736c73e70475f1c6f1f2ee0be81f1c1f53ddb3ab6d15c23b2b2a658a5f205cc","abstract_canon_sha256":"ecb11adedc1c4c5940749e95206607ad0aea9b271f187a3c8a677fbecb012124"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-20T00:05:12.816670Z","signature_b64":"D/SZaZcNetZQ/FeWBJBX0ptpC50yRgiWWjAG408VG3o9M0CISmsqmsVD0UtSYiorNd6v3STtFYgxC8uV3Wu5CA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"9bd6f42859948578f10e86c77193a4080f85796e4ae22c26a20d00ef185b30ec","last_reissued_at":"2026-05-20T00:05:12.815564Z","signature_status":"signed_v1","first_computed_at":"2026-05-20T00:05:12.815564Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"OmniSelect: Dynamic Modality-Aware Token Compression for Efficient Omni-modal Large Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Jianxin Zhang, Juntao Li, Le Li, Morunliu Yang, Peifeng Li, Ruotao Xu, Siwei Feng, Yihang Lou, Yue Wang","submitted_at":"2026-05-18T08:33:52Z","abstract_excerpt":"Omnimodal large language models (OmniLLMs) have recently gained increasing attention for unified audio-video understanding. However, processing long multimodal token sequences introduces substantial computational overhead, making efficient token compression crucial. Existing methods typically rely on fixed, modality-specific guidance, which fails to account for the varying importance of modalities across different queries. To address this limitation, we propose $\\textbf{OmniSelect}$, a training-free, modality-adaptive token pruning framework that dynamically selects appropriate compression str"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.18041","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/2605.18041/integrity.json","findings":[],"available":true,"detectors_run":[{"name":"claim_evidence","ran_at":"2026-05-19T23:41:59.306205Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"ai_meta_artifact","ran_at":"2026-05-19T23:33:35.496766Z","status":"skipped","version":"1.0.0","findings_count":0}],"snapshot_sha256":"5f4d131fc7135cdbf757f8562acee7afb759c2fab172e5eeb8de66c9bb5e0814"},"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":"2605.18041","created_at":"2026-05-20T00:05:12.815906+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.18041v1","created_at":"2026-05-20T00:05:12.815906+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.18041","created_at":"2026-05-20T00:05:12.815906+00:00"},{"alias_kind":"pith_short_12","alias_value":"TPLPIKCZSSCX","created_at":"2026-05-20T00:05:12.815906+00:00"},{"alias_kind":"pith_short_16","alias_value":"TPLPIKCZSSCXR4IO","created_at":"2026-05-20T00:05:12.815906+00:00"},{"alias_kind":"pith_short_8","alias_value":"TPLPIKCZ","created_at":"2026-05-20T00:05:12.815906+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/TPLPIKCZSSCXR4IOQ3DXDE5EBA","json":"https://pith.science/pith/TPLPIKCZSSCXR4IOQ3DXDE5EBA.json","graph_json":"https://pith.science/api/pith-number/TPLPIKCZSSCXR4IOQ3DXDE5EBA/graph.json","events_json":"https://pith.science/api/pith-number/TPLPIKCZSSCXR4IOQ3DXDE5EBA/events.json","paper":"https://pith.science/paper/TPLPIKCZ"},"agent_actions":{"view_html":"https://pith.science/pith/TPLPIKCZSSCXR4IOQ3DXDE5EBA","download_json":"https://pith.science/pith/TPLPIKCZSSCXR4IOQ3DXDE5EBA.json","view_paper":"https://pith.science/paper/TPLPIKCZ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.18041&json=true","fetch_graph":"https://pith.science/api/pith-number/TPLPIKCZSSCXR4IOQ3DXDE5EBA/graph.json","fetch_events":"https://pith.science/api/pith-number/TPLPIKCZSSCXR4IOQ3DXDE5EBA/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/TPLPIKCZSSCXR4IOQ3DXDE5EBA/action/timestamp_anchor","attest_storage":"https://pith.science/pith/TPLPIKCZSSCXR4IOQ3DXDE5EBA/action/storage_attestation","attest_author":"https://pith.science/pith/TPLPIKCZSSCXR4IOQ3DXDE5EBA/action/author_attestation","sign_citation":"https://pith.science/pith/TPLPIKCZSSCXR4IOQ3DXDE5EBA/action/citation_signature","submit_replication":"https://pith.science/pith/TPLPIKCZSSCXR4IOQ3DXDE5EBA/action/replication_record"}},"created_at":"2026-05-20T00:05:12.815906+00:00","updated_at":"2026-05-20T00:05:12.815906+00:00"}