{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:UI7PZNFYMZQ5YF7MJAQNW47MFB","short_pith_number":"pith:UI7PZNFY","schema_version":"1.0","canonical_sha256":"a23efcb4b86661dc17ec4820db73ec2875453b9eec7a48d3a145133329181b24","source":{"kind":"arxiv","id":"2604.00086","version":2},"attestation_state":"computed","paper":{"title":"Hierarchical Pre-Training of Vision Encoders with Large Language Model","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":["cs.AI","cs.CL","cs.LG"],"primary_cat":"cs.CV","authors_text":"Chen-Yi Lee, Eugene Lee, Jiajie Diao, Jui-Huang Tsai, Ting-Yu Chang","submitted_at":"2026-03-31T18:00:39Z","abstract_excerpt":"The field of computer vision has experienced significant advancements through scalable vision encoders and multimodal pre-training frameworks. However, existing approaches often treat vision encoders and large language models (LLMs) as independent modules, limiting the integration of hierarchical visual features. In this work, we propose HIVE (Hierarchical Pre-Training of Vision Encoders), a novel framework that enhances vision-language alignment by introducing hierarchical cross-attention between the vision encoder and LLM. Unlike conventional methods that flatten image embeddings, HIVE enabl"},"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":"2604.00086","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.CV","submitted_at":"2026-03-31T18:00:39Z","cross_cats_sorted":["cs.AI","cs.CL","cs.LG"],"title_canon_sha256":"d2c55fbd6585dc4b6f12a1a72cedb936f64d4313bd6a799bbb975e67c89b652e","abstract_canon_sha256":"aad7a496361870439a9df259f16d11a05abb3da3684bf5222a0d8a5450894768"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-08-04T02:40:30.793901Z","signature_b64":"fOF6iLSgdrNfa2AdBVrE/ndc+iDcoDNb5LvoK/E07+Foz9QpUKFDHecv96qKnzvzmrCcvU+qt9e2rpP5L8i1Cg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"a23efcb4b86661dc17ec4820db73ec2875453b9eec7a48d3a145133329181b24","last_reissued_at":"2026-08-04T02:40:30.791457Z","signature_status":"signed_v1","first_computed_at":"2026-08-04T02:40:30.791457Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Hierarchical Pre-Training of Vision Encoders with Large Language Model","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":["cs.AI","cs.CL","cs.LG"],"primary_cat":"cs.CV","authors_text":"Chen-Yi Lee, Eugene Lee, Jiajie Diao, Jui-Huang Tsai, Ting-Yu Chang","submitted_at":"2026-03-31T18:00:39Z","abstract_excerpt":"The field of computer vision has experienced significant advancements through scalable vision encoders and multimodal pre-training frameworks. However, existing approaches often treat vision encoders and large language models (LLMs) as independent modules, limiting the integration of hierarchical visual features. In this work, we propose HIVE (Hierarchical Pre-Training of Vision Encoders), a novel framework that enhances vision-language alignment by introducing hierarchical cross-attention between the vision encoder and LLM. Unlike conventional methods that flatten image embeddings, HIVE enabl"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2604.00086","kind":"arxiv","version":2},"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/2604.00086/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":"2604.00086","created_at":"2026-08-04T02:40:30.793712+00:00"},{"alias_kind":"arxiv_version","alias_value":"2604.00086v2","created_at":"2026-08-04T02:40:30.793712+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2604.00086","created_at":"2026-08-04T02:40:30.793712+00:00"},{"alias_kind":"pith_short_12","alias_value":"UI7PZNFYMZQ5","created_at":"2026-08-04T02:40:30.793712+00:00"},{"alias_kind":"pith_short_16","alias_value":"UI7PZNFYMZQ5YF7M","created_at":"2026-08-04T02:40:30.793712+00:00"},{"alias_kind":"pith_short_8","alias_value":"UI7PZNFY","created_at":"2026-08-04T02:40:30.793712+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/UI7PZNFYMZQ5YF7MJAQNW47MFB","json":"https://pith.science/pith/UI7PZNFYMZQ5YF7MJAQNW47MFB.json","graph_json":"https://pith.science/api/pith-number/UI7PZNFYMZQ5YF7MJAQNW47MFB/graph.json","events_json":"https://pith.science/api/pith-number/UI7PZNFYMZQ5YF7MJAQNW47MFB/events.json","paper":"https://pith.science/paper/UI7PZNFY"},"agent_actions":{"view_html":"https://pith.science/pith/UI7PZNFYMZQ5YF7MJAQNW47MFB","download_json":"https://pith.science/pith/UI7PZNFYMZQ5YF7MJAQNW47MFB.json","view_paper":"https://pith.science/paper/UI7PZNFY","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2604.00086&json=true","fetch_graph":"https://pith.science/api/pith-number/UI7PZNFYMZQ5YF7MJAQNW47MFB/graph.json","fetch_events":"https://pith.science/api/pith-number/UI7PZNFYMZQ5YF7MJAQNW47MFB/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/UI7PZNFYMZQ5YF7MJAQNW47MFB/action/timestamp_anchor","attest_storage":"https://pith.science/pith/UI7PZNFYMZQ5YF7MJAQNW47MFB/action/storage_attestation","attest_author":"https://pith.science/pith/UI7PZNFYMZQ5YF7MJAQNW47MFB/action/author_attestation","sign_citation":"https://pith.science/pith/UI7PZNFYMZQ5YF7MJAQNW47MFB/action/citation_signature","submit_replication":"https://pith.science/pith/UI7PZNFYMZQ5YF7MJAQNW47MFB/action/replication_record"}},"created_at":"2026-08-04T02:40:30.793712+00:00","updated_at":"2026-08-04T02:40:30.793712+00:00"}