{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:UD72MFPTPDBQ6IRSKMT2WMDIRE","short_pith_number":"pith:UD72MFPT","schema_version":"1.0","canonical_sha256":"a0ffa615f378c30f22325327ab3068890f84a6f407bf6dae375b2dee70837c1a","source":{"kind":"arxiv","id":"2606.04767","version":1},"attestation_state":"computed","paper":{"title":"Measuring Model Robustness via Fisher Information: Spectral Bounds, Theoretical Guarantees, and Practical Algorithms","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CV"],"primary_cat":"cs.LG","authors_text":"Chong Zhang, Jia Wang, Qiufeng Wang, Xiang Li, Xiaobo Jin","submitted_at":"2026-06-03T11:50:38Z","abstract_excerpt":"The robustness of deep neural networks is crucial for safety-critical deployments, yet existing evaluation methods are often attack-dependent and lack interpretability. We propose a principled, attack-agnostic robustness metric based on the spectral norm of the Fisher Information Matrix (FIM), which quantifies the worst-case sensitivity of the model's output distribution to input perturbations. Theoretically, we establish that the FIM equals the variance of the input Jacobian and derive closed-form spectral bounds for common architectures, including VGG, ResNet, DenseNet, and Transformer, prov"},"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":"2606.04767","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-06-03T11:50:38Z","cross_cats_sorted":["cs.CV"],"title_canon_sha256":"1f13dc54d0e3d7b8b27f1df1d58cc849c58a585adfb4a9cfe9e452bc2cdc53d8","abstract_canon_sha256":"3131c548775719bbe7c8d1246ec5173ecf04f452520a8d4415e8ad4c8ca1d739"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-04T01:09:29.244286Z","signature_b64":"MawClFOR1ajyOb9Mnu+VA3DK97sds5A5Xaj5WvbgxKopuGfx1/ecHtZRIuG5ue8T/WDtfjyZdrowcEBR8wi9BA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"a0ffa615f378c30f22325327ab3068890f84a6f407bf6dae375b2dee70837c1a","last_reissued_at":"2026-06-04T01:09:29.243709Z","signature_status":"signed_v1","first_computed_at":"2026-06-04T01:09:29.243709Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Measuring Model Robustness via Fisher Information: Spectral Bounds, Theoretical Guarantees, and Practical Algorithms","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CV"],"primary_cat":"cs.LG","authors_text":"Chong Zhang, Jia Wang, Qiufeng Wang, Xiang Li, Xiaobo Jin","submitted_at":"2026-06-03T11:50:38Z","abstract_excerpt":"The robustness of deep neural networks is crucial for safety-critical deployments, yet existing evaluation methods are often attack-dependent and lack interpretability. We propose a principled, attack-agnostic robustness metric based on the spectral norm of the Fisher Information Matrix (FIM), which quantifies the worst-case sensitivity of the model's output distribution to input perturbations. Theoretically, we establish that the FIM equals the variance of the input Jacobian and derive closed-form spectral bounds for common architectures, including VGG, ResNet, DenseNet, and Transformer, prov"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.04767","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/2606.04767/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":"2606.04767","created_at":"2026-06-04T01:09:29.243809+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.04767v1","created_at":"2026-06-04T01:09:29.243809+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.04767","created_at":"2026-06-04T01:09:29.243809+00:00"},{"alias_kind":"pith_short_12","alias_value":"UD72MFPTPDBQ","created_at":"2026-06-04T01:09:29.243809+00:00"},{"alias_kind":"pith_short_16","alias_value":"UD72MFPTPDBQ6IRS","created_at":"2026-06-04T01:09:29.243809+00:00"},{"alias_kind":"pith_short_8","alias_value":"UD72MFPT","created_at":"2026-06-04T01:09:29.243809+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/UD72MFPTPDBQ6IRSKMT2WMDIRE","json":"https://pith.science/pith/UD72MFPTPDBQ6IRSKMT2WMDIRE.json","graph_json":"https://pith.science/api/pith-number/UD72MFPTPDBQ6IRSKMT2WMDIRE/graph.json","events_json":"https://pith.science/api/pith-number/UD72MFPTPDBQ6IRSKMT2WMDIRE/events.json","paper":"https://pith.science/paper/UD72MFPT"},"agent_actions":{"view_html":"https://pith.science/pith/UD72MFPTPDBQ6IRSKMT2WMDIRE","download_json":"https://pith.science/pith/UD72MFPTPDBQ6IRSKMT2WMDIRE.json","view_paper":"https://pith.science/paper/UD72MFPT","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.04767&json=true","fetch_graph":"https://pith.science/api/pith-number/UD72MFPTPDBQ6IRSKMT2WMDIRE/graph.json","fetch_events":"https://pith.science/api/pith-number/UD72MFPTPDBQ6IRSKMT2WMDIRE/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/UD72MFPTPDBQ6IRSKMT2WMDIRE/action/timestamp_anchor","attest_storage":"https://pith.science/pith/UD72MFPTPDBQ6IRSKMT2WMDIRE/action/storage_attestation","attest_author":"https://pith.science/pith/UD72MFPTPDBQ6IRSKMT2WMDIRE/action/author_attestation","sign_citation":"https://pith.science/pith/UD72MFPTPDBQ6IRSKMT2WMDIRE/action/citation_signature","submit_replication":"https://pith.science/pith/UD72MFPTPDBQ6IRSKMT2WMDIRE/action/replication_record"}},"created_at":"2026-06-04T01:09:29.243809+00:00","updated_at":"2026-06-04T01:09:29.243809+00:00"}