{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:USXFY7WS2PXAGEDHS3JU5JKPCP","short_pith_number":"pith:USXFY7WS","schema_version":"1.0","canonical_sha256":"a4ae5c7ed2d3ee03106796d34ea54f13c825613975d0d8c9f2138ccdbc6a9490","source":{"kind":"arxiv","id":"2507.10095","version":2},"attestation_state":"computed","paper":{"title":"FIX-CLIP: Dual-Branch Hierarchical Contrastive Learning via Synthetic Captions for Better Understanding of Long Text","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Bingchao Wang, Dongsheng Jiang, Jianyu Ding, Jie Yang, Wei Liu, Xuanang Gao, Yin Li, Zhiwei Ning","submitted_at":"2025-07-14T09:31:34Z","abstract_excerpt":"CLIP has shown promising performance across many short-text tasks in a zero-shot manner. However, limited by the input length of the text encoder, CLIP struggles on under-stream tasks with long-text inputs ($>77$ tokens). To remedy this issue, we propose FIX-CLIP, which includes three novel modules: (1) A dual-branch training pipeline that aligns short and long texts with masked and raw images, respectively, which boosts the long-text representation while preserving the short-text ability. (2) Multiple learnable regional prompts with unidirectional masks in Transformer layers for regional info"},"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":"2507.10095","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.CV","submitted_at":"2025-07-14T09:31:34Z","cross_cats_sorted":[],"title_canon_sha256":"55c8e902f5afe3cc2425d3d15a1f87a7975e7dff5feb1800b2da562e32c2cc00","abstract_canon_sha256":"23770703cb012af99ce219902fbc314102469b779ccfe312d4bf540bde7e5168"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:44:51.313933Z","signature_b64":"Eox2M69HvvbA39GZmNTNJ8Sz5Qg+5RGbAG8IMrGnsu6dU+ujJWEHTErfeR9nIaDT0nHqdOwKE/hSa0SefGrlDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"a4ae5c7ed2d3ee03106796d34ea54f13c825613975d0d8c9f2138ccdbc6a9490","last_reissued_at":"2026-07-05T11:44:51.313305Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:44:51.313305Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"FIX-CLIP: Dual-Branch Hierarchical Contrastive Learning via Synthetic Captions for Better Understanding of Long Text","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Bingchao Wang, Dongsheng Jiang, Jianyu Ding, Jie Yang, Wei Liu, Xuanang Gao, Yin Li, Zhiwei Ning","submitted_at":"2025-07-14T09:31:34Z","abstract_excerpt":"CLIP has shown promising performance across many short-text tasks in a zero-shot manner. However, limited by the input length of the text encoder, CLIP struggles on under-stream tasks with long-text inputs ($>77$ tokens). To remedy this issue, we propose FIX-CLIP, which includes three novel modules: (1) A dual-branch training pipeline that aligns short and long texts with masked and raw images, respectively, which boosts the long-text representation while preserving the short-text ability. (2) Multiple learnable regional prompts with unidirectional masks in Transformer layers for regional info"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2507.10095","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/2507.10095/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":"2507.10095","created_at":"2026-07-05T11:44:51.313388+00:00"},{"alias_kind":"arxiv_version","alias_value":"2507.10095v2","created_at":"2026-07-05T11:44:51.313388+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2507.10095","created_at":"2026-07-05T11:44:51.313388+00:00"},{"alias_kind":"pith_short_12","alias_value":"USXFY7WS2PXA","created_at":"2026-07-05T11:44:51.313388+00:00"},{"alias_kind":"pith_short_16","alias_value":"USXFY7WS2PXAGEDH","created_at":"2026-07-05T11:44:51.313388+00:00"},{"alias_kind":"pith_short_8","alias_value":"USXFY7WS","created_at":"2026-07-05T11:44:51.313388+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2508.01219","citing_title":"Eigen Neural Network: Unlocking Generalizable Vision with Eigenbasis","ref_index":42,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/USXFY7WS2PXAGEDHS3JU5JKPCP","json":"https://pith.science/pith/USXFY7WS2PXAGEDHS3JU5JKPCP.json","graph_json":"https://pith.science/api/pith-number/USXFY7WS2PXAGEDHS3JU5JKPCP/graph.json","events_json":"https://pith.science/api/pith-number/USXFY7WS2PXAGEDHS3JU5JKPCP/events.json","paper":"https://pith.science/paper/USXFY7WS"},"agent_actions":{"view_html":"https://pith.science/pith/USXFY7WS2PXAGEDHS3JU5JKPCP","download_json":"https://pith.science/pith/USXFY7WS2PXAGEDHS3JU5JKPCP.json","view_paper":"https://pith.science/paper/USXFY7WS","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2507.10095&json=true","fetch_graph":"https://pith.science/api/pith-number/USXFY7WS2PXAGEDHS3JU5JKPCP/graph.json","fetch_events":"https://pith.science/api/pith-number/USXFY7WS2PXAGEDHS3JU5JKPCP/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/USXFY7WS2PXAGEDHS3JU5JKPCP/action/timestamp_anchor","attest_storage":"https://pith.science/pith/USXFY7WS2PXAGEDHS3JU5JKPCP/action/storage_attestation","attest_author":"https://pith.science/pith/USXFY7WS2PXAGEDHS3JU5JKPCP/action/author_attestation","sign_citation":"https://pith.science/pith/USXFY7WS2PXAGEDHS3JU5JKPCP/action/citation_signature","submit_replication":"https://pith.science/pith/USXFY7WS2PXAGEDHS3JU5JKPCP/action/replication_record"}},"created_at":"2026-07-05T11:44:51.313388+00:00","updated_at":"2026-07-05T11:44:51.313388+00:00"}