{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:Y3QSFFE4ISVIYEY5OCUJKQP2FZ","short_pith_number":"pith:Y3QSFFE4","schema_version":"1.0","canonical_sha256":"c6e122949c44aa8c131d70a89541fa2e67c5acc74741d4b9a07ae161aad1f097","source":{"kind":"arxiv","id":"2608.07088","version":1},"attestation_state":"computed","paper":{"title":"RoRA: Role-Oriented Regional Allocation for Visual Token Pruning in MLLMs","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Cheng Fan, Han Wu, Jianyuan Guo, Jufeng Yang, Minjing Dong, Qiyanhui Lu, Rongjian Xu, Tingzhang Luo, Xinghao Chen","submitted_at":"2026-08-07T10:39:47Z","abstract_excerpt":"Multimodal large language models (MLLMs) encode images as long visual token sequences, making prefilling and KV-cache storage expensive. Existing training-free pruning methods select tokens by importance, diversity, or spatial coverage, but treat retained tokens as interchangeable and do not explicitly track which object-related regions are already covered. We present RoRA, a training-free framework that casts visual token pruning as role-oriented regional evidence allocation. Given a fixed budget, RoRA partitions tokens into a protected semantic core, complementary context, and fine-grained d"},"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":"2608.07088","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2026-08-07T10:39:47Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"8590c43f70420a466ff56afa230ca1f1454b2a59dc9976b2b55403fad7845770","abstract_canon_sha256":"8d7721ce94907c53a87e589249755c3a70c90513af86a1b7b92aa5c9e6877bde"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-08-10T01:13:11.230609Z","signature_b64":"EHAJLADEEnT1e7Yxe3N+EGclgnBpZiliwVv+ULMiW8XdCPd4NLbhgdvjGEYjj87wcT3hHvWbxoEUjhKJmF9iBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"c6e122949c44aa8c131d70a89541fa2e67c5acc74741d4b9a07ae161aad1f097","last_reissued_at":"2026-08-10T01:13:11.227879Z","signature_status":"signed_v1","first_computed_at":"2026-08-10T01:13:11.227879Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"RoRA: Role-Oriented Regional Allocation for Visual Token Pruning in MLLMs","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Cheng Fan, Han Wu, Jianyuan Guo, Jufeng Yang, Minjing Dong, Qiyanhui Lu, Rongjian Xu, Tingzhang Luo, Xinghao Chen","submitted_at":"2026-08-07T10:39:47Z","abstract_excerpt":"Multimodal large language models (MLLMs) encode images as long visual token sequences, making prefilling and KV-cache storage expensive. Existing training-free pruning methods select tokens by importance, diversity, or spatial coverage, but treat retained tokens as interchangeable and do not explicitly track which object-related regions are already covered. We present RoRA, a training-free framework that casts visual token pruning as role-oriented regional evidence allocation. Given a fixed budget, RoRA partitions tokens into a protected semantic core, complementary context, and fine-grained d"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2608.07088","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/2608.07088/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":"2608.07088","created_at":"2026-08-10T01:13:11.228958+00:00"},{"alias_kind":"arxiv_version","alias_value":"2608.07088v1","created_at":"2026-08-10T01:13:11.228958+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2608.07088","created_at":"2026-08-10T01:13:11.228958+00:00"},{"alias_kind":"pith_short_12","alias_value":"Y3QSFFE4ISVI","created_at":"2026-08-10T01:13:11.228958+00:00"},{"alias_kind":"pith_short_16","alias_value":"Y3QSFFE4ISVIYEY5","created_at":"2026-08-10T01:13:11.228958+00:00"},{"alias_kind":"pith_short_8","alias_value":"Y3QSFFE4","created_at":"2026-08-10T01:13:11.228958+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/Y3QSFFE4ISVIYEY5OCUJKQP2FZ","json":"https://pith.science/pith/Y3QSFFE4ISVIYEY5OCUJKQP2FZ.json","graph_json":"https://pith.science/api/pith-number/Y3QSFFE4ISVIYEY5OCUJKQP2FZ/graph.json","events_json":"https://pith.science/api/pith-number/Y3QSFFE4ISVIYEY5OCUJKQP2FZ/events.json","paper":"https://pith.science/paper/Y3QSFFE4"},"agent_actions":{"view_html":"https://pith.science/pith/Y3QSFFE4ISVIYEY5OCUJKQP2FZ","download_json":"https://pith.science/pith/Y3QSFFE4ISVIYEY5OCUJKQP2FZ.json","view_paper":"https://pith.science/paper/Y3QSFFE4","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2608.07088&json=true","fetch_graph":"https://pith.science/api/pith-number/Y3QSFFE4ISVIYEY5OCUJKQP2FZ/graph.json","fetch_events":"https://pith.science/api/pith-number/Y3QSFFE4ISVIYEY5OCUJKQP2FZ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/Y3QSFFE4ISVIYEY5OCUJKQP2FZ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/Y3QSFFE4ISVIYEY5OCUJKQP2FZ/action/storage_attestation","attest_author":"https://pith.science/pith/Y3QSFFE4ISVIYEY5OCUJKQP2FZ/action/author_attestation","sign_citation":"https://pith.science/pith/Y3QSFFE4ISVIYEY5OCUJKQP2FZ/action/citation_signature","submit_replication":"https://pith.science/pith/Y3QSFFE4ISVIYEY5OCUJKQP2FZ/action/replication_record"}},"created_at":"2026-08-10T01:13:11.228958+00:00","updated_at":"2026-08-10T01:13:11.228958+00:00"}