{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:PVIGMX6VPNLPMRTR4OSPNPSOVF","short_pith_number":"pith:PVIGMX6V","schema_version":"1.0","canonical_sha256":"7d50665fd57b56f64671e3a4f6be4ea95b01b7a11081b1e8123d29438fcf2595","source":{"kind":"arxiv","id":"2409.01179","version":3},"attestation_state":"computed","paper":{"title":"Recoverable Compression: A Multimodal Vision Token Recovery Mechanism Guided by Text Information","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Cheng-Lin Liu, Jian Xu, Wen-Zhuo Liu, Xu-Yao Zhang, Yang-Yang Liu, Yi Chen","submitted_at":"2024-09-02T11:19:54Z","abstract_excerpt":"With the advancement of large-scale language modeling techniques, large multimodal models combining visual encoders with large language models have demonstrated exceptional performance in various visual tasks. Most of the current large-scale multimodal models achieve this by mapping visual features obtained from the visual encoder into a large language model and using them as inputs alongside text for downstream tasks. Therefore, the number of visual tokens directly affects the training and inference speed of the model. There has been significant work on token pruning for visual transformers, "},"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":"2409.01179","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2024-09-02T11:19:54Z","cross_cats_sorted":[],"title_canon_sha256":"287927a6e25a4ef8e7988ce116a8f831a7b8a1123d3e9989f77f70a283f5e5cf","abstract_canon_sha256":"d42e802bb362522d6172cd4cfb6255d628d2378f1477d70aca16a7ebf3909b32"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:51:36.206621Z","signature_b64":"PpXcejplPYDh+IswSjk0UaqjJMQU2T6Qs6cpoRCo5G6g9l95oBShAjAzvWtHjS1SPZL91/dDgjTjQ+sx6fSADg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"7d50665fd57b56f64671e3a4f6be4ea95b01b7a11081b1e8123d29438fcf2595","last_reissued_at":"2026-07-05T09:51:36.206156Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:51:36.206156Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Recoverable Compression: A Multimodal Vision Token Recovery Mechanism Guided by Text Information","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Cheng-Lin Liu, Jian Xu, Wen-Zhuo Liu, Xu-Yao Zhang, Yang-Yang Liu, Yi Chen","submitted_at":"2024-09-02T11:19:54Z","abstract_excerpt":"With the advancement of large-scale language modeling techniques, large multimodal models combining visual encoders with large language models have demonstrated exceptional performance in various visual tasks. Most of the current large-scale multimodal models achieve this by mapping visual features obtained from the visual encoder into a large language model and using them as inputs alongside text for downstream tasks. Therefore, the number of visual tokens directly affects the training and inference speed of the model. There has been significant work on token pruning for visual transformers, "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2409.01179","kind":"arxiv","version":3},"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/2409.01179/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":"2409.01179","created_at":"2026-07-05T09:51:36.206212+00:00"},{"alias_kind":"arxiv_version","alias_value":"2409.01179v3","created_at":"2026-07-05T09:51:36.206212+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2409.01179","created_at":"2026-07-05T09:51:36.206212+00:00"},{"alias_kind":"pith_short_12","alias_value":"PVIGMX6VPNLP","created_at":"2026-07-05T09:51:36.206212+00:00"},{"alias_kind":"pith_short_16","alias_value":"PVIGMX6VPNLPMRTR","created_at":"2026-07-05T09:51:36.206212+00:00"},{"alias_kind":"pith_short_8","alias_value":"PVIGMX6V","created_at":"2026-07-05T09:51:36.206212+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2604.19167","citing_title":"LBLLM: Lightweight Binarization of Large Language Models via Three-Stage Distillation","ref_index":70,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/PVIGMX6VPNLPMRTR4OSPNPSOVF","json":"https://pith.science/pith/PVIGMX6VPNLPMRTR4OSPNPSOVF.json","graph_json":"https://pith.science/api/pith-number/PVIGMX6VPNLPMRTR4OSPNPSOVF/graph.json","events_json":"https://pith.science/api/pith-number/PVIGMX6VPNLPMRTR4OSPNPSOVF/events.json","paper":"https://pith.science/paper/PVIGMX6V"},"agent_actions":{"view_html":"https://pith.science/pith/PVIGMX6VPNLPMRTR4OSPNPSOVF","download_json":"https://pith.science/pith/PVIGMX6VPNLPMRTR4OSPNPSOVF.json","view_paper":"https://pith.science/paper/PVIGMX6V","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2409.01179&json=true","fetch_graph":"https://pith.science/api/pith-number/PVIGMX6VPNLPMRTR4OSPNPSOVF/graph.json","fetch_events":"https://pith.science/api/pith-number/PVIGMX6VPNLPMRTR4OSPNPSOVF/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/PVIGMX6VPNLPMRTR4OSPNPSOVF/action/timestamp_anchor","attest_storage":"https://pith.science/pith/PVIGMX6VPNLPMRTR4OSPNPSOVF/action/storage_attestation","attest_author":"https://pith.science/pith/PVIGMX6VPNLPMRTR4OSPNPSOVF/action/author_attestation","sign_citation":"https://pith.science/pith/PVIGMX6VPNLPMRTR4OSPNPSOVF/action/citation_signature","submit_replication":"https://pith.science/pith/PVIGMX6VPNLPMRTR4OSPNPSOVF/action/replication_record"}},"created_at":"2026-07-05T09:51:36.206212+00:00","updated_at":"2026-07-05T09:51:36.206212+00:00"}