{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:USETYIEIX4223A6DJFUDUUZU3Y","short_pith_number":"pith:USETYIEI","schema_version":"1.0","canonical_sha256":"a4893c2088bf35ad83c349683a5334de3db4316a4d314ba33de1b38737152192","source":{"kind":"arxiv","id":"2602.04804","version":2},"attestation_state":"computed","paper":{"title":"OmniSIFT: Modality-Asymmetric Token Compression for Efficient Omni-modal Large Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"OmniSIFT reduces omni-modal token sequences to 25 percent of their length while matching or exceeding full-context accuracy on several tasks.","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Bohan Zeng, Bozhou Li, Jiaheng Liu, Junfei Wu, Jungang Li, Liang Wang, Pengfei Wan, Qiang Liu, Xinlong Chen, Xuyang Liu, Yang Shi, Yiyan Ji, Yuanxing Zhang, Yue Ding, Yushuo Guan","submitted_at":"2026-02-04T17:51:05Z","abstract_excerpt":"Omni-modal Large Language Models (Omni-LLMs) have demonstrated strong capabilities in audio-video understanding tasks. However, their reliance on long multimodal token sequences leads to substantial computational overhead. Despite this challenge, token compression methods designed for Omni-LLMs remain limited. To bridge this gap, we propose OmniSIFT (Omni-modal Spatio-temporal Informed Fine-grained Token compression), a modality-asymmetric token compression framework tailored for Omni-LLMs. Specifically, OmniSIFT adopts a two-stage compression strategy: (i) a spatio-temporal video pruning modu"},"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":"2602.04804","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2026-02-04T17:51:05Z","cross_cats_sorted":[],"title_canon_sha256":"1b810817ebc3d1fb0e4f4d8c781cb53a299cbaf5deaa2f2ad26b2e117ea283df","abstract_canon_sha256":"752f62d099b3fa5a36b0b92e0ff005a1cc6de75a9612c0a1c333663bd761c5a8"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-18T02:44:31.636159Z","signature_b64":"SYbMikyR9d7frcjMd8cv9wZOff8uBKVv28aUIa9knAD3/9ac7KJ8HMRaduF6jFHZaXi7Kx4dlJlQgLpxgqEbAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"a4893c2088bf35ad83c349683a5334de3db4316a4d314ba33de1b38737152192","last_reissued_at":"2026-05-18T02:44:31.635686Z","signature_status":"signed_v1","first_computed_at":"2026-05-18T02:44:31.635686Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"OmniSIFT: Modality-Asymmetric Token Compression for Efficient Omni-modal Large Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"OmniSIFT reduces omni-modal token sequences to 25 percent of their length while matching or exceeding full-context accuracy on several tasks.","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Bohan Zeng, Bozhou Li, Jiaheng Liu, Junfei Wu, Jungang Li, Liang Wang, Pengfei Wan, Qiang Liu, Xinlong Chen, Xuyang Liu, Yang Shi, Yiyan Ji, Yuanxing Zhang, Yue Ding, Yushuo Guan","submitted_at":"2026-02-04T17:51:05Z","abstract_excerpt":"Omni-modal Large Language Models (Omni-LLMs) have demonstrated strong capabilities in audio-video understanding tasks. However, their reliance on long multimodal token sequences leads to substantial computational overhead. Despite this challenge, token compression methods designed for Omni-LLMs remain limited. To bridge this gap, we propose OmniSIFT (Omni-modal Spatio-temporal Informed Fine-grained Token compression), a modality-asymmetric token compression framework tailored for Omni-LLMs. Specifically, OmniSIFT adopts a two-stage compression strategy: (i) a spatio-temporal video pruning modu"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"With merely 25% of the original token context, OmniSIFT consistently outperforms all compression baselines and even surpasses the performance of the full-token model on several tasks.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That the spatio-temporal pruning and vision-guided audio selection preserve all task-critical information across diverse video-audio content without introducing systematic bias.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"OmniSIFT delivers modality-asymmetric token compression for omni-LLMs via spatio-temporal video pruning and vision-guided audio selection, outperforming baselines at 25% token count.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"OmniSIFT reduces omni-modal token sequences to 25 percent of their length while matching or exceeding full-context accuracy on several tasks.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"f84376d72f1a270cb1af84fc693fa689699909ab1be784bb1ed8f4afd857eea7"},"source":{"id":"2602.04804","kind":"arxiv","version":2},"verdict":{"id":"1c2b82ee-18a6-486d-8527-6019016f34fe","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-16T07:21:07.168044Z","strongest_claim":"With merely 25% of the original token context, OmniSIFT consistently outperforms all compression baselines and even surpasses the performance of the full-token model on several tasks.","one_line_summary":"OmniSIFT delivers modality-asymmetric token compression for omni-LLMs via spatio-temporal video pruning and vision-guided audio selection, outperforming baselines at 25% token count.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That the spatio-temporal pruning and vision-guided audio selection preserve all task-critical information across diverse video-audio content without introducing systematic bias.","pith_extraction_headline":"OmniSIFT reduces omni-modal token sequences to 25 percent of their length while matching or exceeding full-context accuracy on several tasks."},"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":"2602.04804","created_at":"2026-05-18T02:44:31.635758+00:00"},{"alias_kind":"arxiv_version","alias_value":"2602.04804v2","created_at":"2026-05-18T02:44:31.635758+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2602.04804","created_at":"2026-05-18T02:44:31.635758+00:00"},{"alias_kind":"pith_short_12","alias_value":"USETYIEIX422","created_at":"2026-05-18T12:33:37.589309+00:00"},{"alias_kind":"pith_short_16","alias_value":"USETYIEIX4223A6D","created_at":"2026-05-18T12:33:37.589309+00:00"},{"alias_kind":"pith_short_8","alias_value":"USETYIEI","created_at":"2026-05-18T12:33:37.589309+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":7,"internal_anchor_count":7,"sample":[{"citing_arxiv_id":"2605.22012","citing_title":"LatentOmni: Rethinking Omni-Modal Understanding via Unified Audio-Visual Latent Reasoning","ref_index":11,"is_internal_anchor":true},{"citing_arxiv_id":"2605.20035","citing_title":"Stage-adaptive Token Selection for Efficient Omni-modal LLMs","ref_index":5,"is_internal_anchor":true},{"citing_arxiv_id":"2605.12056","citing_title":"OmniRefine: Alignment-Aware Cooperative Compression for Efficient Omnimodal Large Language Models","ref_index":12,"is_internal_anchor":true},{"citing_arxiv_id":"2605.11605","citing_title":"Keep What Audio Cannot Say: Context-Preserving Token Pruning for Omni-LLMs","ref_index":13,"is_internal_anchor":true},{"citing_arxiv_id":"2604.23717","citing_title":"HeadRouter: Dynamic Head-Weight Routing for Task-Adaptive Audio Token Pruning in Large Audio Language Models","ref_index":7,"is_internal_anchor":true},{"citing_arxiv_id":"2605.07593","citing_title":"TraceAV-Bench: Benchmarking Multi-Hop Trajectory Reasoning over Long Audio-Visual Videos","ref_index":52,"is_internal_anchor":true},{"citing_arxiv_id":"2604.04707","citing_title":"OpenWorldLib: A Unified Codebase and Definition of Advanced World Models","ref_index":27,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/USETYIEIX4223A6DJFUDUUZU3Y","json":"https://pith.science/pith/USETYIEIX4223A6DJFUDUUZU3Y.json","graph_json":"https://pith.science/api/pith-number/USETYIEIX4223A6DJFUDUUZU3Y/graph.json","events_json":"https://pith.science/api/pith-number/USETYIEIX4223A6DJFUDUUZU3Y/events.json","paper":"https://pith.science/paper/USETYIEI"},"agent_actions":{"view_html":"https://pith.science/pith/USETYIEIX4223A6DJFUDUUZU3Y","download_json":"https://pith.science/pith/USETYIEIX4223A6DJFUDUUZU3Y.json","view_paper":"https://pith.science/paper/USETYIEI","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2602.04804&json=true","fetch_graph":"https://pith.science/api/pith-number/USETYIEIX4223A6DJFUDUUZU3Y/graph.json","fetch_events":"https://pith.science/api/pith-number/USETYIEIX4223A6DJFUDUUZU3Y/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/USETYIEIX4223A6DJFUDUUZU3Y/action/timestamp_anchor","attest_storage":"https://pith.science/pith/USETYIEIX4223A6DJFUDUUZU3Y/action/storage_attestation","attest_author":"https://pith.science/pith/USETYIEIX4223A6DJFUDUUZU3Y/action/author_attestation","sign_citation":"https://pith.science/pith/USETYIEIX4223A6DJFUDUUZU3Y/action/citation_signature","submit_replication":"https://pith.science/pith/USETYIEIX4223A6DJFUDUUZU3Y/action/replication_record"}},"created_at":"2026-05-18T02:44:31.635758+00:00","updated_at":"2026-05-18T02:44:31.635758+00:00"}