{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:XGYKGKELLB27GWS6QI64VWFZST","short_pith_number":"pith:XGYKGKEL","schema_version":"1.0","canonical_sha256":"b9b0a3288b5875f35a5e823dcad8b994d37db6e64763d978c8deab016db7563a","source":{"kind":"arxiv","id":"2605.04209","version":2},"attestation_state":"computed","paper":{"title":"Undetectable Backdoors in Model Parameters: Hiding Sparse Secrets in High Dimensions","license":"http://creativecommons.org/licenses/by/4.0/","headline":"A sparse perturbation masked by Gaussian dither embeds backdoors whose detection reduces to the hard Sparse PCA problem.","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CR","authors_text":"Ashish Hooda, Atharv Singh Patlan, Kassem Fawaz, Nils Palumbo, Sarthak Choudhary, Somesh Jha","submitted_at":"2026-05-05T18:48:09Z","abstract_excerpt":"We present Sparse Backdoor, a supply-chain attack that plants a provably undetectable backdoor in pre-trained image classifiers, including convolutional networks and Vision Transformers. The attack injects a structured sparse perturbation along a randomly chosen direction into a small subset of columns at each fully connected layer, propagating a trigger signal to an adversary-chosen target class, and masks the perturbation with an independent isotropic Gaussian dither. The dither serves a single technical purpose: it induces a clean reference distribution anchored at the pre-trained weights, "},"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":true},"canonical_record":{"source":{"id":"2605.04209","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CR","submitted_at":"2026-05-05T18:48:09Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"0fa9e7dc09eafd9c787eb6c712b03eb3824dfafa953e4cac6c0d93b4ed0489b0","abstract_canon_sha256":"be79287cd0b4835d87a4ea0069326fbe1bcf2588d839ce266ab0d2c61886c46d"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-07T02:17:26.416176Z","signature_b64":"SS3sJCoOg2pkNMdYrw6PGFT0SJ+gUFdpkqh2CECPa4/j9KSHaN+M4RrATEN0kSyi3wbpwXN9UnuG+jxBTzyRDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"b9b0a3288b5875f35a5e823dcad8b994d37db6e64763d978c8deab016db7563a","last_reissued_at":"2026-07-07T02:17:26.415350Z","signature_status":"signed_v1","first_computed_at":"2026-07-07T02:17:26.415350Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Undetectable Backdoors in Model Parameters: Hiding Sparse Secrets in High Dimensions","license":"http://creativecommons.org/licenses/by/4.0/","headline":"A sparse perturbation masked by Gaussian dither embeds backdoors whose detection reduces to the hard Sparse PCA problem.","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CR","authors_text":"Ashish Hooda, Atharv Singh Patlan, Kassem Fawaz, Nils Palumbo, Sarthak Choudhary, Somesh Jha","submitted_at":"2026-05-05T18:48:09Z","abstract_excerpt":"We present Sparse Backdoor, a supply-chain attack that plants a provably undetectable backdoor in pre-trained image classifiers, including convolutional networks and Vision Transformers. The attack injects a structured sparse perturbation along a randomly chosen direction into a small subset of columns at each fully connected layer, propagating a trigger signal to an adversary-chosen target class, and masks the perturbation with an independent isotropic Gaussian dither. The dither serves a single technical purpose: it induces a clean reference distribution anchored at the pre-trained weights, "},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"We prove that distinguishing the backdoor-injected model from this reference is at least as hard as Sparse PCA detection, which is computationally infeasible under standard hardness assumptions. The guarantee holds against any probabilistic polynomial-time distinguisher with white-box access to the parameters.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"Under a mild margin condition on the pre-trained classifier, we show that the dithered reference is functionally equivalent to the original classifier.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"Sparse Backdoor plants a provably undetectable backdoor in neural network weights via structured sparse perturbations and isotropic Gaussian dithering, with detection hardness reduced to Sparse PCA.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"A sparse perturbation masked by Gaussian dither embeds backdoors whose detection reduces to the hard Sparse PCA problem.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"4abf5a0048e9af099c8f815670b13a585c52cba3679462b0359f0ec9967571ba"},"source":{"id":"2605.04209","kind":"arxiv","version":2},"verdict":{"id":"d5f04a65-83af-45f4-ace9-2a199959f977","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-08T17:46:52.224990Z","strongest_claim":"We prove that distinguishing the backdoor-injected model from this reference is at least as hard as Sparse PCA detection, which is computationally infeasible under standard hardness assumptions. The guarantee holds against any probabilistic polynomial-time distinguisher with white-box access to the parameters.","one_line_summary":"Sparse Backdoor plants a provably undetectable backdoor in neural network weights via structured sparse perturbations and isotropic Gaussian dithering, with detection hardness reduced to Sparse PCA.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"Under a mild margin condition on the pre-trained classifier, we show that the dithered reference is functionally equivalent to the original classifier.","pith_extraction_headline":"A sparse perturbation masked by Gaussian dither embeds backdoors whose detection reduces to the hard Sparse PCA problem."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2605.04209/integrity.json","findings":[],"available":true,"detectors_run":[{"name":"ai_meta_artifact","ran_at":"2026-05-20T12:36:29.502660Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"doi_title_agreement","ran_at":"2026-05-19T23:31:21.106278Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"doi_compliance","ran_at":"2026-05-19T14:42:12.068385Z","status":"completed","version":"1.0.0","findings_count":0}],"snapshot_sha256":"8bfe0ea30b0c8efd3dea90410b7bf352ab53d4d5b4b3b87a22774ab37e819647"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":3,"snapshot_sha256":"2da0535982af8a43209d79cace1004967f5ad8af61d552cf59704771c860b437"},"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":"2605.04209","created_at":"2026-07-07T02:17:26.415480+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.04209v2","created_at":"2026-07-07T02:17:26.415480+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.04209","created_at":"2026-07-07T02:17:26.415480+00:00"},{"alias_kind":"pith_short_12","alias_value":"XGYKGKELLB27","created_at":"2026-07-07T02:17:26.415480+00:00"},{"alias_kind":"pith_short_16","alias_value":"XGYKGKELLB27GWS6","created_at":"2026-07-07T02:17:26.415480+00:00"},{"alias_kind":"pith_short_8","alias_value":"XGYKGKEL","created_at":"2026-07-07T02:17:26.415480+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2607.09532","citing_title":"Statistically Undetectable Backdoors in Deep Neural Networks","ref_index":58,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":3,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/XGYKGKELLB27GWS6QI64VWFZST","json":"https://pith.science/pith/XGYKGKELLB27GWS6QI64VWFZST.json","graph_json":"https://pith.science/api/pith-number/XGYKGKELLB27GWS6QI64VWFZST/graph.json","events_json":"https://pith.science/api/pith-number/XGYKGKELLB27GWS6QI64VWFZST/events.json","paper":"https://pith.science/paper/XGYKGKEL"},"agent_actions":{"view_html":"https://pith.science/pith/XGYKGKELLB27GWS6QI64VWFZST","download_json":"https://pith.science/pith/XGYKGKELLB27GWS6QI64VWFZST.json","view_paper":"https://pith.science/paper/XGYKGKEL","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.04209&json=true","fetch_graph":"https://pith.science/api/pith-number/XGYKGKELLB27GWS6QI64VWFZST/graph.json","fetch_events":"https://pith.science/api/pith-number/XGYKGKELLB27GWS6QI64VWFZST/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/XGYKGKELLB27GWS6QI64VWFZST/action/timestamp_anchor","attest_storage":"https://pith.science/pith/XGYKGKELLB27GWS6QI64VWFZST/action/storage_attestation","attest_author":"https://pith.science/pith/XGYKGKELLB27GWS6QI64VWFZST/action/author_attestation","sign_citation":"https://pith.science/pith/XGYKGKELLB27GWS6QI64VWFZST/action/citation_signature","submit_replication":"https://pith.science/pith/XGYKGKELLB27GWS6QI64VWFZST/action/replication_record"}},"created_at":"2026-07-07T02:17:26.415480+00:00","updated_at":"2026-07-07T02:17:26.415480+00:00"}