{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:7GHXM65S7N2QIM52KSMC6R3HS6","short_pith_number":"pith:7GHXM65S","schema_version":"1.0","canonical_sha256":"f98f767bb2fb750433ba54982f4767979e13a11d66cae657b84c69a0a33d5203","source":{"kind":"arxiv","id":"2103.15670","version":3},"attestation_state":"computed","paper":{"title":"On the Adversarial Robustness of Vision Transformers","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CV","authors_text":"Cho-Jui Hsieh, Jinfeng Yi, Pin-Yu Chen, Rulin Shao, Zhouxing Shi","submitted_at":"2021-03-29T14:48:24Z","abstract_excerpt":"Following the success in advancing natural language processing and understanding, transformers are expected to bring revolutionary changes to computer vision. This work provides a comprehensive study on the robustness of vision transformers (ViTs) against adversarial perturbations. Tested on various white-box and transfer attack settings, we find that ViTs possess better adversarial robustness when compared with MLP-Mixer and convolutional neural networks (CNNs) including ConvNeXt, and this observation also holds for certified robustness. Through frequency analysis and feature visualization, w"},"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":"2103.15670","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2021-03-29T14:48:24Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"b07ae2ed6c1c2a2c02b5307137c555f99cd67a716b9b1773e174435325b352a5","abstract_canon_sha256":"375d5c3af65fcca9fcaf00b0b80090467a84ebf12ff1deb346dd352faac7de20"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T05:12:47.999348Z","signature_b64":"oaQbZUnxHxQg5w8LKcOhM9if20qFQdx6gR38RzCRBpWtza/i5cwAELT7zBzFccvNYB5632wyjvvmz0OTQGUaCw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"f98f767bb2fb750433ba54982f4767979e13a11d66cae657b84c69a0a33d5203","last_reissued_at":"2026-07-05T05:12:47.998855Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T05:12:47.998855Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"On the Adversarial Robustness of Vision Transformers","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CV","authors_text":"Cho-Jui Hsieh, Jinfeng Yi, Pin-Yu Chen, Rulin Shao, Zhouxing Shi","submitted_at":"2021-03-29T14:48:24Z","abstract_excerpt":"Following the success in advancing natural language processing and understanding, transformers are expected to bring revolutionary changes to computer vision. This work provides a comprehensive study on the robustness of vision transformers (ViTs) against adversarial perturbations. Tested on various white-box and transfer attack settings, we find that ViTs possess better adversarial robustness when compared with MLP-Mixer and convolutional neural networks (CNNs) including ConvNeXt, and this observation also holds for certified robustness. Through frequency analysis and feature visualization, w"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2103.15670","kind":"arxiv","version":3},"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/2103.15670/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":"2103.15670","created_at":"2026-07-05T05:12:47.998913+00:00"},{"alias_kind":"arxiv_version","alias_value":"2103.15670v3","created_at":"2026-07-05T05:12:47.998913+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2103.15670","created_at":"2026-07-05T05:12:47.998913+00:00"},{"alias_kind":"pith_short_12","alias_value":"7GHXM65S7N2Q","created_at":"2026-07-05T05:12:47.998913+00:00"},{"alias_kind":"pith_short_16","alias_value":"7GHXM65S7N2QIM52","created_at":"2026-07-05T05:12:47.998913+00:00"},{"alias_kind":"pith_short_8","alias_value":"7GHXM65S","created_at":"2026-07-05T05:12:47.998913+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":3,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.00738","citing_title":"SORA: Free Second-Order Attacks in Fast Adversarial Training","ref_index":14,"is_internal_anchor":false},{"citing_arxiv_id":"2605.18868","citing_title":"DarkLLM: Learning Language-Driven Adversarial Attacks with Large Language Models","ref_index":48,"is_internal_anchor":false},{"citing_arxiv_id":"2604.19724","citing_title":"Benign Overfitting in Adversarial Training for Vision Transformers","ref_index":29,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/7GHXM65S7N2QIM52KSMC6R3HS6","json":"https://pith.science/pith/7GHXM65S7N2QIM52KSMC6R3HS6.json","graph_json":"https://pith.science/api/pith-number/7GHXM65S7N2QIM52KSMC6R3HS6/graph.json","events_json":"https://pith.science/api/pith-number/7GHXM65S7N2QIM52KSMC6R3HS6/events.json","paper":"https://pith.science/paper/7GHXM65S"},"agent_actions":{"view_html":"https://pith.science/pith/7GHXM65S7N2QIM52KSMC6R3HS6","download_json":"https://pith.science/pith/7GHXM65S7N2QIM52KSMC6R3HS6.json","view_paper":"https://pith.science/paper/7GHXM65S","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2103.15670&json=true","fetch_graph":"https://pith.science/api/pith-number/7GHXM65S7N2QIM52KSMC6R3HS6/graph.json","fetch_events":"https://pith.science/api/pith-number/7GHXM65S7N2QIM52KSMC6R3HS6/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/7GHXM65S7N2QIM52KSMC6R3HS6/action/timestamp_anchor","attest_storage":"https://pith.science/pith/7GHXM65S7N2QIM52KSMC6R3HS6/action/storage_attestation","attest_author":"https://pith.science/pith/7GHXM65S7N2QIM52KSMC6R3HS6/action/author_attestation","sign_citation":"https://pith.science/pith/7GHXM65S7N2QIM52KSMC6R3HS6/action/citation_signature","submit_replication":"https://pith.science/pith/7GHXM65S7N2QIM52KSMC6R3HS6/action/replication_record"}},"created_at":"2026-07-05T05:12:47.998913+00:00","updated_at":"2026-07-05T05:12:47.998913+00:00"}