{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:QRZBX2FR3FVU7S673YVC5ANM2L","short_pith_number":"pith:QRZBX2FR","schema_version":"1.0","canonical_sha256":"84721be8b1d96b4fcbdfde2a2e81acd2d79a7c4ff47480fbd9bdad38f94c9453","source":{"kind":"arxiv","id":"2605.28115","version":1},"attestation_state":"computed","paper":{"title":"CIVIC: End-to-End Sequence Compactness for Efficient Vision-Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Bo Yu, Cathy Liu, Chenxi Liu, Fengze Yang, Xuewen Luo","submitted_at":"2026-05-27T08:09:44Z","abstract_excerpt":"Vision-Language Models (VLMs) face severe memory and latency bottlenecks due to high-resolution visual tokens. While current token reduction methods theoretically save FLOPs, post-hoc pruning introduces structural overhead, failing to yield proportional wall-clock acceleration. However, enforcing a contiguous compact pathway risks geometric disorientation and loss of fine-grained localization. To overcome these barriers, this paper introduces CIVIC, a path-consistent compact visual inference framework. By maintaining compact sequence representations seamlessly across the vision encoder, projec"},"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":"2605.28115","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2026-05-27T08:09:44Z","cross_cats_sorted":[],"title_canon_sha256":"14f7f53d3a39b89b30df5502baeaaf0f1df1230f96988e0b63f445285d32c614","abstract_canon_sha256":"efefe3c0f404ce8e58f9df55245fc45476e018f9fd69544949b69abc8d0e2763"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-28T01:04:59.325958Z","signature_b64":"40iuT6SAzuL0WcrS4d3vF/jUnPpnBJ3YGMACOXg4p23V1JxqblLAgV2kWHgmtGqPHFkNoOzJY0Eo9etatsWoDg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"84721be8b1d96b4fcbdfde2a2e81acd2d79a7c4ff47480fbd9bdad38f94c9453","last_reissued_at":"2026-05-28T01:04:59.325450Z","signature_status":"signed_v1","first_computed_at":"2026-05-28T01:04:59.325450Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"CIVIC: End-to-End Sequence Compactness for Efficient Vision-Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Bo Yu, Cathy Liu, Chenxi Liu, Fengze Yang, Xuewen Luo","submitted_at":"2026-05-27T08:09:44Z","abstract_excerpt":"Vision-Language Models (VLMs) face severe memory and latency bottlenecks due to high-resolution visual tokens. While current token reduction methods theoretically save FLOPs, post-hoc pruning introduces structural overhead, failing to yield proportional wall-clock acceleration. However, enforcing a contiguous compact pathway risks geometric disorientation and loss of fine-grained localization. To overcome these barriers, this paper introduces CIVIC, a path-consistent compact visual inference framework. By maintaining compact sequence representations seamlessly across the vision encoder, projec"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.28115","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/2605.28115/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":"2605.28115","created_at":"2026-05-28T01:04:59.325529+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.28115v1","created_at":"2026-05-28T01:04:59.325529+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.28115","created_at":"2026-05-28T01:04:59.325529+00:00"},{"alias_kind":"pith_short_12","alias_value":"QRZBX2FR3FVU","created_at":"2026-05-28T01:04:59.325529+00:00"},{"alias_kind":"pith_short_16","alias_value":"QRZBX2FR3FVU7S67","created_at":"2026-05-28T01:04:59.325529+00:00"},{"alias_kind":"pith_short_8","alias_value":"QRZBX2FR","created_at":"2026-05-28T01:04:59.325529+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/QRZBX2FR3FVU7S673YVC5ANM2L","json":"https://pith.science/pith/QRZBX2FR3FVU7S673YVC5ANM2L.json","graph_json":"https://pith.science/api/pith-number/QRZBX2FR3FVU7S673YVC5ANM2L/graph.json","events_json":"https://pith.science/api/pith-number/QRZBX2FR3FVU7S673YVC5ANM2L/events.json","paper":"https://pith.science/paper/QRZBX2FR"},"agent_actions":{"view_html":"https://pith.science/pith/QRZBX2FR3FVU7S673YVC5ANM2L","download_json":"https://pith.science/pith/QRZBX2FR3FVU7S673YVC5ANM2L.json","view_paper":"https://pith.science/paper/QRZBX2FR","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.28115&json=true","fetch_graph":"https://pith.science/api/pith-number/QRZBX2FR3FVU7S673YVC5ANM2L/graph.json","fetch_events":"https://pith.science/api/pith-number/QRZBX2FR3FVU7S673YVC5ANM2L/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/QRZBX2FR3FVU7S673YVC5ANM2L/action/timestamp_anchor","attest_storage":"https://pith.science/pith/QRZBX2FR3FVU7S673YVC5ANM2L/action/storage_attestation","attest_author":"https://pith.science/pith/QRZBX2FR3FVU7S673YVC5ANM2L/action/author_attestation","sign_citation":"https://pith.science/pith/QRZBX2FR3FVU7S673YVC5ANM2L/action/citation_signature","submit_replication":"https://pith.science/pith/QRZBX2FR3FVU7S673YVC5ANM2L/action/replication_record"}},"created_at":"2026-05-28T01:04:59.325529+00:00","updated_at":"2026-05-28T01:04:59.325529+00:00"}