{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:JJMXHVWZ2O62D2TRWMHCTGYZBN","short_pith_number":"pith:JJMXHVWZ","schema_version":"1.0","canonical_sha256":"4a5973d6d9d3bda1ea71b30e299b190b622520635d27f7b287bb62d6040a9b87","source":{"kind":"arxiv","id":"2606.31383","version":1},"attestation_state":"computed","paper":{"title":"MS-Resampler: Multi-Scope Visual Resampling for Efficient Multimodal LLMs","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Cheng Qian, Faming Fang, Guixu Zhang, Kaiwen Long, Mo Guang, Rinyoichi Takezoe, Yaqian Li, Zhongyang Li, Zi-Hao Bo","submitted_at":"2026-06-30T09:11:50Z","abstract_excerpt":"Multimodal large language models (MLLMs) typically employ resampling-based projectors to transform dense visual features into a compact token sequence for language modeling. Most existing resamplers adopt a single, fixed aggregation scope via global cross-attention, which can blur fine-grained local evidence and limit the ability to capture both local details and global context within a fixed token budget. In this work, we propose MS-Resampler, a multi-scope visual resampling framework for MLLMs. MS-Resampler instantiates multiple scope-specific resamplers by injecting explicit spatial scope p"},"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":"2606.31383","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2026-06-30T09:11:50Z","cross_cats_sorted":[],"title_canon_sha256":"fc0cc9c40e556c0afa01328c522aff907c05766674c92836c11932c50f8ab401","abstract_canon_sha256":"af478cfcaeefa9cc3a944b03f9338b4ec6dfd9334d454ec0c69192fe6209400c"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-01T01:18:01.440519Z","signature_b64":"hfXR5inIfgU8fO1yUC8gViXdx9ax6s1Mjt7dc643YmZZqfIrXR1Ofkw1C7cYCNYfpDvHgkfUdTrNlHixQMVEBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"4a5973d6d9d3bda1ea71b30e299b190b622520635d27f7b287bb62d6040a9b87","last_reissued_at":"2026-07-01T01:18:01.440082Z","signature_status":"signed_v1","first_computed_at":"2026-07-01T01:18:01.440082Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"MS-Resampler: Multi-Scope Visual Resampling for Efficient Multimodal LLMs","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Cheng Qian, Faming Fang, Guixu Zhang, Kaiwen Long, Mo Guang, Rinyoichi Takezoe, Yaqian Li, Zhongyang Li, Zi-Hao Bo","submitted_at":"2026-06-30T09:11:50Z","abstract_excerpt":"Multimodal large language models (MLLMs) typically employ resampling-based projectors to transform dense visual features into a compact token sequence for language modeling. Most existing resamplers adopt a single, fixed aggregation scope via global cross-attention, which can blur fine-grained local evidence and limit the ability to capture both local details and global context within a fixed token budget. In this work, we propose MS-Resampler, a multi-scope visual resampling framework for MLLMs. MS-Resampler instantiates multiple scope-specific resamplers by injecting explicit spatial scope p"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.31383","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/2606.31383/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":"2606.31383","created_at":"2026-07-01T01:18:01.440139+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.31383v1","created_at":"2026-07-01T01:18:01.440139+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.31383","created_at":"2026-07-01T01:18:01.440139+00:00"},{"alias_kind":"pith_short_12","alias_value":"JJMXHVWZ2O62","created_at":"2026-07-01T01:18:01.440139+00:00"},{"alias_kind":"pith_short_16","alias_value":"JJMXHVWZ2O62D2TR","created_at":"2026-07-01T01:18:01.440139+00:00"},{"alias_kind":"pith_short_8","alias_value":"JJMXHVWZ","created_at":"2026-07-01T01:18:01.440139+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/JJMXHVWZ2O62D2TRWMHCTGYZBN","json":"https://pith.science/pith/JJMXHVWZ2O62D2TRWMHCTGYZBN.json","graph_json":"https://pith.science/api/pith-number/JJMXHVWZ2O62D2TRWMHCTGYZBN/graph.json","events_json":"https://pith.science/api/pith-number/JJMXHVWZ2O62D2TRWMHCTGYZBN/events.json","paper":"https://pith.science/paper/JJMXHVWZ"},"agent_actions":{"view_html":"https://pith.science/pith/JJMXHVWZ2O62D2TRWMHCTGYZBN","download_json":"https://pith.science/pith/JJMXHVWZ2O62D2TRWMHCTGYZBN.json","view_paper":"https://pith.science/paper/JJMXHVWZ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.31383&json=true","fetch_graph":"https://pith.science/api/pith-number/JJMXHVWZ2O62D2TRWMHCTGYZBN/graph.json","fetch_events":"https://pith.science/api/pith-number/JJMXHVWZ2O62D2TRWMHCTGYZBN/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/JJMXHVWZ2O62D2TRWMHCTGYZBN/action/timestamp_anchor","attest_storage":"https://pith.science/pith/JJMXHVWZ2O62D2TRWMHCTGYZBN/action/storage_attestation","attest_author":"https://pith.science/pith/JJMXHVWZ2O62D2TRWMHCTGYZBN/action/author_attestation","sign_citation":"https://pith.science/pith/JJMXHVWZ2O62D2TRWMHCTGYZBN/action/citation_signature","submit_replication":"https://pith.science/pith/JJMXHVWZ2O62D2TRWMHCTGYZBN/action/replication_record"}},"created_at":"2026-07-01T01:18:01.440139+00:00","updated_at":"2026-07-01T01:18:01.440139+00:00"}