{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:JL3BO5LLUTNQWB2OYSUM44RVWG","short_pith_number":"pith:JL3BO5LL","schema_version":"1.0","canonical_sha256":"4af617756ba4db0b074ec4a8ce7235b1aece22bfc16c0d930da0ae7a70f7ed88","source":{"kind":"arxiv","id":"2606.18839","version":1},"attestation_state":"computed","paper":{"title":"Semantic Robustness Certification for Vision-Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CV"],"primary_cat":"cs.LG","authors_text":"Amardeep Kaur, Andrew C. Cullen, Christopher Leckie, Feng Liu, Paul Montague, Peiyu Yang, Sarah M. Erfani","submitted_at":"2026-06-17T09:15:50Z","abstract_excerpt":"Vision-language models (VLMs) are now widely used in downstream tasks. However, real-world applications often expose VLMs to distribution shifts induced by semantic variation (e.g., shape, size, and style). Robustness certification determines if a model's prediction changes when transformations are applied to its input. While most certification frameworks study geometric or pixel-level transformations over inputs, this work proposes a novel framework that enables certifying VLM robustness under semantic-level transformations. Leveraging the open-vocabulary capability of VLMs, we use text promp"},"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":"2606.18839","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-06-17T09:15:50Z","cross_cats_sorted":["cs.CV"],"title_canon_sha256":"d75e242311196dec115b002b6d2e5884c08a1d43cf6c3d5d1c3488eb04df6121","abstract_canon_sha256":"9c4152fef94733f5d0aa91bec8202daba039f2571d488ca5acf744e417c22026"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-19T16:11:48.963413Z","signature_b64":"q7m+lPoiTeThX74pc5zR94ikHobk4TDryBbM89CJOHCnZJJj925QcNCA0cHyZ/i8ho1JulYwOGUPn0aGT0bvAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"4af617756ba4db0b074ec4a8ce7235b1aece22bfc16c0d930da0ae7a70f7ed88","last_reissued_at":"2026-06-19T16:11:48.963030Z","signature_status":"signed_v1","first_computed_at":"2026-06-19T16:11:48.963030Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Semantic Robustness Certification for Vision-Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CV"],"primary_cat":"cs.LG","authors_text":"Amardeep Kaur, Andrew C. Cullen, Christopher Leckie, Feng Liu, Paul Montague, Peiyu Yang, Sarah M. Erfani","submitted_at":"2026-06-17T09:15:50Z","abstract_excerpt":"Vision-language models (VLMs) are now widely used in downstream tasks. However, real-world applications often expose VLMs to distribution shifts induced by semantic variation (e.g., shape, size, and style). Robustness certification determines if a model's prediction changes when transformations are applied to its input. While most certification frameworks study geometric or pixel-level transformations over inputs, this work proposes a novel framework that enables certifying VLM robustness under semantic-level transformations. Leveraging the open-vocabulary capability of VLMs, we use text promp"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.18839","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/2606.18839/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":"2606.18839","created_at":"2026-06-19T16:11:48.963092+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.18839v1","created_at":"2026-06-19T16:11:48.963092+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.18839","created_at":"2026-06-19T16:11:48.963092+00:00"},{"alias_kind":"pith_short_12","alias_value":"JL3BO5LLUTNQ","created_at":"2026-06-19T16:11:48.963092+00:00"},{"alias_kind":"pith_short_16","alias_value":"JL3BO5LLUTNQWB2O","created_at":"2026-06-19T16:11:48.963092+00:00"},{"alias_kind":"pith_short_8","alias_value":"JL3BO5LL","created_at":"2026-06-19T16:11:48.963092+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/JL3BO5LLUTNQWB2OYSUM44RVWG","json":"https://pith.science/pith/JL3BO5LLUTNQWB2OYSUM44RVWG.json","graph_json":"https://pith.science/api/pith-number/JL3BO5LLUTNQWB2OYSUM44RVWG/graph.json","events_json":"https://pith.science/api/pith-number/JL3BO5LLUTNQWB2OYSUM44RVWG/events.json","paper":"https://pith.science/paper/JL3BO5LL"},"agent_actions":{"view_html":"https://pith.science/pith/JL3BO5LLUTNQWB2OYSUM44RVWG","download_json":"https://pith.science/pith/JL3BO5LLUTNQWB2OYSUM44RVWG.json","view_paper":"https://pith.science/paper/JL3BO5LL","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.18839&json=true","fetch_graph":"https://pith.science/api/pith-number/JL3BO5LLUTNQWB2OYSUM44RVWG/graph.json","fetch_events":"https://pith.science/api/pith-number/JL3BO5LLUTNQWB2OYSUM44RVWG/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/JL3BO5LLUTNQWB2OYSUM44RVWG/action/timestamp_anchor","attest_storage":"https://pith.science/pith/JL3BO5LLUTNQWB2OYSUM44RVWG/action/storage_attestation","attest_author":"https://pith.science/pith/JL3BO5LLUTNQWB2OYSUM44RVWG/action/author_attestation","sign_citation":"https://pith.science/pith/JL3BO5LLUTNQWB2OYSUM44RVWG/action/citation_signature","submit_replication":"https://pith.science/pith/JL3BO5LLUTNQWB2OYSUM44RVWG/action/replication_record"}},"created_at":"2026-06-19T16:11:48.963092+00:00","updated_at":"2026-06-19T16:11:48.963092+00:00"}