{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:MCLKZ7GC6LFXVXJITQ7Q523XCE","short_pith_number":"pith:MCLKZ7GC","schema_version":"1.0","canonical_sha256":"6096acfcc2f2cb7add289c3f0eeb771102bccf4a56f118fb274ebd78c931201a","source":{"kind":"arxiv","id":"2606.01710","version":1},"attestation_state":"computed","paper":{"title":"Density-Aware Translation of Spurious Correlations in Zero-Shot VLMs","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CV","authors_text":"Afsaneh Hasanebrahimi, Christopher Leckie, Hanxun Huang, Sarah Erfani","submitted_at":"2026-06-01T05:23:48Z","abstract_excerpt":"Vision-Language models (VLMs), such as CLIP, achieve powerful zero-shot classification. However, their predictions remain sensitive to spurious correlations, where contextual cues dominate over semantic content. Earlier solutions typically rely on fine-tuning or prompt engineering, which either undermine the advantages of pre-trained models or are prone to hallucination. In this work, we propose Density-Aware Translation (DAT) that refines image-text similarity scores using a local geometric density term derived from group reference sets. Our approach is motivated by the phenomenon that CLIP e"},"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":"2606.01710","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2026-06-01T05:23:48Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"9907791dae04ee15143e273d657e778a256bbae6c553cdcc13d3af33b3948eb6","abstract_canon_sha256":"1b22260439c617ebd6df8d5614589d6f7e435785e3c0ed4e9e903fe16ffed05e"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-02T02:04:40.605128Z","signature_b64":"LbqlqVG6x6ic74prv4+NT1SPxiU4EMQ2yJWw3yOOm6n3LinlRkA3kK1GVg5t2n5kNEFPJZZRuGzRGcj+O6MrDw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"6096acfcc2f2cb7add289c3f0eeb771102bccf4a56f118fb274ebd78c931201a","last_reissued_at":"2026-06-02T02:04:40.604797Z","signature_status":"signed_v1","first_computed_at":"2026-06-02T02:04:40.604797Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Density-Aware Translation of Spurious Correlations in Zero-Shot VLMs","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CV","authors_text":"Afsaneh Hasanebrahimi, Christopher Leckie, Hanxun Huang, Sarah Erfani","submitted_at":"2026-06-01T05:23:48Z","abstract_excerpt":"Vision-Language models (VLMs), such as CLIP, achieve powerful zero-shot classification. However, their predictions remain sensitive to spurious correlations, where contextual cues dominate over semantic content. Earlier solutions typically rely on fine-tuning or prompt engineering, which either undermine the advantages of pre-trained models or are prone to hallucination. In this work, we propose Density-Aware Translation (DAT) that refines image-text similarity scores using a local geometric density term derived from group reference sets. Our approach is motivated by the phenomenon that CLIP e"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.01710","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/2606.01710/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":"2606.01710","created_at":"2026-06-02T02:04:40.604851+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.01710v1","created_at":"2026-06-02T02:04:40.604851+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.01710","created_at":"2026-06-02T02:04:40.604851+00:00"},{"alias_kind":"pith_short_12","alias_value":"MCLKZ7GC6LFX","created_at":"2026-06-02T02:04:40.604851+00:00"},{"alias_kind":"pith_short_16","alias_value":"MCLKZ7GC6LFXVXJI","created_at":"2026-06-02T02:04:40.604851+00:00"},{"alias_kind":"pith_short_8","alias_value":"MCLKZ7GC","created_at":"2026-06-02T02:04:40.604851+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/MCLKZ7GC6LFXVXJITQ7Q523XCE","json":"https://pith.science/pith/MCLKZ7GC6LFXVXJITQ7Q523XCE.json","graph_json":"https://pith.science/api/pith-number/MCLKZ7GC6LFXVXJITQ7Q523XCE/graph.json","events_json":"https://pith.science/api/pith-number/MCLKZ7GC6LFXVXJITQ7Q523XCE/events.json","paper":"https://pith.science/paper/MCLKZ7GC"},"agent_actions":{"view_html":"https://pith.science/pith/MCLKZ7GC6LFXVXJITQ7Q523XCE","download_json":"https://pith.science/pith/MCLKZ7GC6LFXVXJITQ7Q523XCE.json","view_paper":"https://pith.science/paper/MCLKZ7GC","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.01710&json=true","fetch_graph":"https://pith.science/api/pith-number/MCLKZ7GC6LFXVXJITQ7Q523XCE/graph.json","fetch_events":"https://pith.science/api/pith-number/MCLKZ7GC6LFXVXJITQ7Q523XCE/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/MCLKZ7GC6LFXVXJITQ7Q523XCE/action/timestamp_anchor","attest_storage":"https://pith.science/pith/MCLKZ7GC6LFXVXJITQ7Q523XCE/action/storage_attestation","attest_author":"https://pith.science/pith/MCLKZ7GC6LFXVXJITQ7Q523XCE/action/author_attestation","sign_citation":"https://pith.science/pith/MCLKZ7GC6LFXVXJITQ7Q523XCE/action/citation_signature","submit_replication":"https://pith.science/pith/MCLKZ7GC6LFXVXJITQ7Q523XCE/action/replication_record"}},"created_at":"2026-06-02T02:04:40.604851+00:00","updated_at":"2026-06-02T02:04:40.604851+00:00"}