{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:BVLSC4HLQ6BCJGSZD3RFQUGHGO","short_pith_number":"pith:BVLSC4HL","schema_version":"1.0","canonical_sha256":"0d572170eb8782249a591ee25850c73387f7036fef45f6c8d50716dd4f5bfdd8","source":{"kind":"arxiv","id":"2607.02484","version":1},"attestation_state":"computed","paper":{"title":"Combating Textual Noise and Redundancy: Entropy-Aware Dense Visual Token Pruning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Wei Shen, Xuankun Yang, Xuehui Wang","submitted_at":"2026-07-02T17:50:57Z","abstract_excerpt":"Visual token pruning is a crucial strategy for accelerating VLMs by compressing redundant image patches, yet existing methods often fail to preserve critical cues under dense instructions and fine-grained queries. In this paper, we investigate this failure and identify two underlying bottlenecks: the widespread dispersion of textual noise that corrupts dense cross-modal scoring, and the feature fragmentation inherent to standard token selection. To address these issues, we propose Entropy-Aware Dense Pruning (EADP), a framework that reformulates pruning as a structured compression problem. EAD"},"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":"2607.02484","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2026-07-02T17:50:57Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"3293961127195ca59b937ad4505a9a8cd8925d8ac7e3079a42836c6743761a10","abstract_canon_sha256":"6aabd10c96db960e91ed616b587676531b67297dc83f3cbfd6b48cdc5a461afe"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-03T01:18:00.507157Z","signature_b64":"I7vWWxe3xBcVqaIP5y2tmGh3dYXAtuDym5egGHCfmqECDQHDCLkKK+TOrYVtcC7BuBXKPRyX7klzbcvmBqVTAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"0d572170eb8782249a591ee25850c73387f7036fef45f6c8d50716dd4f5bfdd8","last_reissued_at":"2026-07-03T01:18:00.506742Z","signature_status":"signed_v1","first_computed_at":"2026-07-03T01:18:00.506742Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Combating Textual Noise and Redundancy: Entropy-Aware Dense Visual Token Pruning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Wei Shen, Xuankun Yang, Xuehui Wang","submitted_at":"2026-07-02T17:50:57Z","abstract_excerpt":"Visual token pruning is a crucial strategy for accelerating VLMs by compressing redundant image patches, yet existing methods often fail to preserve critical cues under dense instructions and fine-grained queries. In this paper, we investigate this failure and identify two underlying bottlenecks: the widespread dispersion of textual noise that corrupts dense cross-modal scoring, and the feature fragmentation inherent to standard token selection. To address these issues, we propose Entropy-Aware Dense Pruning (EADP), a framework that reformulates pruning as a structured compression problem. EAD"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2607.02484","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/2607.02484/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":"2607.02484","created_at":"2026-07-03T01:18:00.506808+00:00"},{"alias_kind":"arxiv_version","alias_value":"2607.02484v1","created_at":"2026-07-03T01:18:00.506808+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2607.02484","created_at":"2026-07-03T01:18:00.506808+00:00"},{"alias_kind":"pith_short_12","alias_value":"BVLSC4HLQ6BC","created_at":"2026-07-03T01:18:00.506808+00:00"},{"alias_kind":"pith_short_16","alias_value":"BVLSC4HLQ6BCJGSZ","created_at":"2026-07-03T01:18:00.506808+00:00"},{"alias_kind":"pith_short_8","alias_value":"BVLSC4HL","created_at":"2026-07-03T01:18:00.506808+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/BVLSC4HLQ6BCJGSZD3RFQUGHGO","json":"https://pith.science/pith/BVLSC4HLQ6BCJGSZD3RFQUGHGO.json","graph_json":"https://pith.science/api/pith-number/BVLSC4HLQ6BCJGSZD3RFQUGHGO/graph.json","events_json":"https://pith.science/api/pith-number/BVLSC4HLQ6BCJGSZD3RFQUGHGO/events.json","paper":"https://pith.science/paper/BVLSC4HL"},"agent_actions":{"view_html":"https://pith.science/pith/BVLSC4HLQ6BCJGSZD3RFQUGHGO","download_json":"https://pith.science/pith/BVLSC4HLQ6BCJGSZD3RFQUGHGO.json","view_paper":"https://pith.science/paper/BVLSC4HL","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2607.02484&json=true","fetch_graph":"https://pith.science/api/pith-number/BVLSC4HLQ6BCJGSZD3RFQUGHGO/graph.json","fetch_events":"https://pith.science/api/pith-number/BVLSC4HLQ6BCJGSZD3RFQUGHGO/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/BVLSC4HLQ6BCJGSZD3RFQUGHGO/action/timestamp_anchor","attest_storage":"https://pith.science/pith/BVLSC4HLQ6BCJGSZD3RFQUGHGO/action/storage_attestation","attest_author":"https://pith.science/pith/BVLSC4HLQ6BCJGSZD3RFQUGHGO/action/author_attestation","sign_citation":"https://pith.science/pith/BVLSC4HLQ6BCJGSZD3RFQUGHGO/action/citation_signature","submit_replication":"https://pith.science/pith/BVLSC4HLQ6BCJGSZD3RFQUGHGO/action/replication_record"}},"created_at":"2026-07-03T01:18:00.506808+00:00","updated_at":"2026-07-03T01:18:00.506808+00:00"}