{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2022:JXIAMKZJEOBLEDTKJGANOUGBWO","short_pith_number":"pith:JXIAMKZJ","schema_version":"1.0","canonical_sha256":"4dd0062b292382b20e6a4980d750c1b383c82538658498abe69db362cc575a71","source":{"kind":"arxiv","id":"2201.07012","version":1},"attestation_state":"computed","paper":{"title":"Adversarial vulnerability of powerful near out-of-distribution detection","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Stanislav Fort","submitted_at":"2022-01-18T14:23:07Z","abstract_excerpt":"There has been a significant progress in detecting out-of-distribution (OOD) inputs in neural networks recently, primarily due to the use of large models pretrained on large datasets, and an emerging use of multi-modality. We show a severe adversarial vulnerability of even the strongest current OOD detection techniques. With a small, targeted perturbation to the input pixels, we can change the image assignment from an in-distribution to an out-distribution, and vice versa, easily. In particular, we demonstrate severe adversarial vulnerability on the challenging near OOD CIFAR-100 vs CIFAR-10 t"},"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":"2201.07012","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2022-01-18T14:23:07Z","cross_cats_sorted":[],"title_canon_sha256":"e6fbbd14f9167d5eeda53bc42a62aaef7826c4206474380130edff5dba85e9a1","abstract_canon_sha256":"1857ad79bb957655a0382d07c37938251b9f10a5d08c9ceef7d828dafbd3f570"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T03:49:33.632206Z","signature_b64":"2agwi9OIQ0dYn1hgGKsEkF7/GAHH0kk/gr4la8ToHmNgt4fLmA/yzsimtXkgKLfz2b/g8OvcQPiRgW6em9lACA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"4dd0062b292382b20e6a4980d750c1b383c82538658498abe69db362cc575a71","last_reissued_at":"2026-07-05T03:49:33.631818Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T03:49:33.631818Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Adversarial vulnerability of powerful near out-of-distribution detection","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Stanislav Fort","submitted_at":"2022-01-18T14:23:07Z","abstract_excerpt":"There has been a significant progress in detecting out-of-distribution (OOD) inputs in neural networks recently, primarily due to the use of large models pretrained on large datasets, and an emerging use of multi-modality. We show a severe adversarial vulnerability of even the strongest current OOD detection techniques. With a small, targeted perturbation to the input pixels, we can change the image assignment from an in-distribution to an out-distribution, and vice versa, easily. In particular, we demonstrate severe adversarial vulnerability on the challenging near OOD CIFAR-100 vs CIFAR-10 t"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2201.07012","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/2201.07012/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":"2201.07012","created_at":"2026-07-05T03:49:33.631875+00:00"},{"alias_kind":"arxiv_version","alias_value":"2201.07012v1","created_at":"2026-07-05T03:49:33.631875+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2201.07012","created_at":"2026-07-05T03:49:33.631875+00:00"},{"alias_kind":"pith_short_12","alias_value":"JXIAMKZJEOBL","created_at":"2026-07-05T03:49:33.631875+00:00"},{"alias_kind":"pith_short_16","alias_value":"JXIAMKZJEOBLEDTK","created_at":"2026-07-05T03:49:33.631875+00:00"},{"alias_kind":"pith_short_8","alias_value":"JXIAMKZJ","created_at":"2026-07-05T03:49:33.631875+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2502.07753","citing_title":"Direct Ascent Synthesis: Revealing Hidden Generative Capabilities in Discriminative Models","ref_index":12,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/JXIAMKZJEOBLEDTKJGANOUGBWO","json":"https://pith.science/pith/JXIAMKZJEOBLEDTKJGANOUGBWO.json","graph_json":"https://pith.science/api/pith-number/JXIAMKZJEOBLEDTKJGANOUGBWO/graph.json","events_json":"https://pith.science/api/pith-number/JXIAMKZJEOBLEDTKJGANOUGBWO/events.json","paper":"https://pith.science/paper/JXIAMKZJ"},"agent_actions":{"view_html":"https://pith.science/pith/JXIAMKZJEOBLEDTKJGANOUGBWO","download_json":"https://pith.science/pith/JXIAMKZJEOBLEDTKJGANOUGBWO.json","view_paper":"https://pith.science/paper/JXIAMKZJ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2201.07012&json=true","fetch_graph":"https://pith.science/api/pith-number/JXIAMKZJEOBLEDTKJGANOUGBWO/graph.json","fetch_events":"https://pith.science/api/pith-number/JXIAMKZJEOBLEDTKJGANOUGBWO/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/JXIAMKZJEOBLEDTKJGANOUGBWO/action/timestamp_anchor","attest_storage":"https://pith.science/pith/JXIAMKZJEOBLEDTKJGANOUGBWO/action/storage_attestation","attest_author":"https://pith.science/pith/JXIAMKZJEOBLEDTKJGANOUGBWO/action/author_attestation","sign_citation":"https://pith.science/pith/JXIAMKZJEOBLEDTKJGANOUGBWO/action/citation_signature","submit_replication":"https://pith.science/pith/JXIAMKZJEOBLEDTKJGANOUGBWO/action/replication_record"}},"created_at":"2026-07-05T03:49:33.631875+00:00","updated_at":"2026-07-05T03:49:33.631875+00:00"}