{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:4YNUIARFLOW6SQMCZOLPV42ZQ2","short_pith_number":"pith:4YNUIARF","schema_version":"1.0","canonical_sha256":"e61b4402255bade94182cb96faf3598688c3376447cb175fec7d9ccb275a72fe","source":{"kind":"arxiv","id":"2603.23483","version":2},"attestation_state":"computed","paper":{"title":"SpecEyes: Accelerating Agentic Multimodal LLMs via Speculative Perception and Planning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.CV","authors_text":"Haoyu Huang, Jiebo Luo, Jinfa Huang, Rongrong Ji, Xiawu Zheng, Zhongwei Wan","submitted_at":"2026-03-24T17:45:47Z","abstract_excerpt":"Agentic multimodal large language models (MLLMs) (e.g., OpenAI o3 and Gemini Agentic Vision) achieve remarkable reasoning capabilities through iterative visual tool invocation. However, the cascaded perception, reasoning, and tool-calling loops introduce significant sequential overhead. This overhead, termed agentic depth, incurs prohibitive latency and seriously limits system-level concurrency. To this end, we propose SpecEyes, an agentic-level speculative acceleration framework that breaks this sequential bottleneck. Our key insight is that a lightweight, tool-free MLLM can serve as a specul"},"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":"2603.23483","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2026-03-24T17:45:47Z","cross_cats_sorted":["cs.CL"],"title_canon_sha256":"1413722383a5d589eae74210985b89051aad11f1bd80336b70f5e947c94a7d1c","abstract_canon_sha256":"88834ed3addfcab06d624cf6118639e8ace81d0d238238849fdf19b034483180"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-07T02:19:49.905864Z","signature_b64":"pn38jqjDjTgBJUIDs8wLd5JVJhyLE8BbAggslLtyKvzRd/f2H2Bxuy85EOir4MTmB4Oye0TAFRUNygsbIgb3AQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e61b4402255bade94182cb96faf3598688c3376447cb175fec7d9ccb275a72fe","last_reissued_at":"2026-07-07T02:19:49.904679Z","signature_status":"signed_v1","first_computed_at":"2026-07-07T02:19:49.904679Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"SpecEyes: Accelerating Agentic Multimodal LLMs via Speculative Perception and Planning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.CV","authors_text":"Haoyu Huang, Jiebo Luo, Jinfa Huang, Rongrong Ji, Xiawu Zheng, Zhongwei Wan","submitted_at":"2026-03-24T17:45:47Z","abstract_excerpt":"Agentic multimodal large language models (MLLMs) (e.g., OpenAI o3 and Gemini Agentic Vision) achieve remarkable reasoning capabilities through iterative visual tool invocation. However, the cascaded perception, reasoning, and tool-calling loops introduce significant sequential overhead. This overhead, termed agentic depth, incurs prohibitive latency and seriously limits system-level concurrency. To this end, we propose SpecEyes, an agentic-level speculative acceleration framework that breaks this sequential bottleneck. Our key insight is that a lightweight, tool-free MLLM can serve as a specul"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2603.23483","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/2603.23483/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":"2603.23483","created_at":"2026-07-07T02:19:49.904826+00:00"},{"alias_kind":"arxiv_version","alias_value":"2603.23483v2","created_at":"2026-07-07T02:19:49.904826+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2603.23483","created_at":"2026-07-07T02:19:49.904826+00:00"},{"alias_kind":"pith_short_12","alias_value":"4YNUIARFLOW6","created_at":"2026-07-07T02:19:49.904826+00:00"},{"alias_kind":"pith_short_16","alias_value":"4YNUIARFLOW6SQMC","created_at":"2026-07-07T02:19:49.904826+00:00"},{"alias_kind":"pith_short_8","alias_value":"4YNUIARF","created_at":"2026-07-07T02:19:49.904826+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/4YNUIARFLOW6SQMCZOLPV42ZQ2","json":"https://pith.science/pith/4YNUIARFLOW6SQMCZOLPV42ZQ2.json","graph_json":"https://pith.science/api/pith-number/4YNUIARFLOW6SQMCZOLPV42ZQ2/graph.json","events_json":"https://pith.science/api/pith-number/4YNUIARFLOW6SQMCZOLPV42ZQ2/events.json","paper":"https://pith.science/paper/4YNUIARF"},"agent_actions":{"view_html":"https://pith.science/pith/4YNUIARFLOW6SQMCZOLPV42ZQ2","download_json":"https://pith.science/pith/4YNUIARFLOW6SQMCZOLPV42ZQ2.json","view_paper":"https://pith.science/paper/4YNUIARF","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2603.23483&json=true","fetch_graph":"https://pith.science/api/pith-number/4YNUIARFLOW6SQMCZOLPV42ZQ2/graph.json","fetch_events":"https://pith.science/api/pith-number/4YNUIARFLOW6SQMCZOLPV42ZQ2/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/4YNUIARFLOW6SQMCZOLPV42ZQ2/action/timestamp_anchor","attest_storage":"https://pith.science/pith/4YNUIARFLOW6SQMCZOLPV42ZQ2/action/storage_attestation","attest_author":"https://pith.science/pith/4YNUIARFLOW6SQMCZOLPV42ZQ2/action/author_attestation","sign_citation":"https://pith.science/pith/4YNUIARFLOW6SQMCZOLPV42ZQ2/action/citation_signature","submit_replication":"https://pith.science/pith/4YNUIARFLOW6SQMCZOLPV42ZQ2/action/replication_record"}},"created_at":"2026-07-07T02:19:49.904826+00:00","updated_at":"2026-07-07T02:19:49.904826+00:00"}