{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:VK2PEP62OCYWXHMY4S5D6VQC2P","short_pith_number":"pith:VK2PEP62","schema_version":"1.0","canonical_sha256":"aab4f23fda70b16b9d98e4ba3f5602d3ed96c06c95f02aabd60dc191811614fc","source":{"kind":"arxiv","id":"2106.08361","version":1},"attestation_state":"computed","paper":{"title":"Adversarial Attacks on Deep Models for Financial Transaction Records","license":"http://creativecommons.org/publicdomain/zero/1.0/","headline":"","cross_cats":["cs.CR","q-fin.ST"],"primary_cat":"cs.LG","authors_text":"Alexey Zaytsev, Dmitry Babaev, Elizaveta Kovtun, Evgeny Burnaev, Gleb Gusev, Ivan Fursov, Ivan Kireev, Matvey Morozov, Nina Kaploukhaya, Rodrigo Rivera-Castro","submitted_at":"2021-06-15T18:15:26Z","abstract_excerpt":"Machine learning models using transaction records as inputs are popular among financial institutions. The most efficient models use deep-learning architectures similar to those in the NLP community, posing a challenge due to their tremendous number of parameters and limited robustness. In particular, deep-learning models are vulnerable to adversarial attacks: a little change in the input harms the model's output.\n  In this work, we examine adversarial attacks on transaction records data and defences from these attacks. The transaction records data have a different structure than the canonical "},"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":"2106.08361","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/publicdomain/zero/1.0/","primary_cat":"cs.LG","submitted_at":"2021-06-15T18:15:26Z","cross_cats_sorted":["cs.CR","q-fin.ST"],"title_canon_sha256":"fdbe152dea3bcbfb0d1f8ee620fa65c3ce224d68965773b474cb60971dab941e","abstract_canon_sha256":"d347333d47133837d5bf0031e3532bba0941432e70e4b8cf03a4e65d80644394"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T02:49:57.029682Z","signature_b64":"1A9DDaTmHw0zsSowW/A4bJPaj9YSdkLfcNFINtL0ggmazSsmSutv/QqGRm47O6ffoRANvL5clMijkoOydHoLAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"aab4f23fda70b16b9d98e4ba3f5602d3ed96c06c95f02aabd60dc191811614fc","last_reissued_at":"2026-07-05T02:49:57.029164Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T02:49:57.029164Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Adversarial Attacks on Deep Models for Financial Transaction Records","license":"http://creativecommons.org/publicdomain/zero/1.0/","headline":"","cross_cats":["cs.CR","q-fin.ST"],"primary_cat":"cs.LG","authors_text":"Alexey Zaytsev, Dmitry Babaev, Elizaveta Kovtun, Evgeny Burnaev, Gleb Gusev, Ivan Fursov, Ivan Kireev, Matvey Morozov, Nina Kaploukhaya, Rodrigo Rivera-Castro","submitted_at":"2021-06-15T18:15:26Z","abstract_excerpt":"Machine learning models using transaction records as inputs are popular among financial institutions. The most efficient models use deep-learning architectures similar to those in the NLP community, posing a challenge due to their tremendous number of parameters and limited robustness. In particular, deep-learning models are vulnerable to adversarial attacks: a little change in the input harms the model's output.\n  In this work, we examine adversarial attacks on transaction records data and defences from these attacks. The transaction records data have a different structure than the canonical "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2106.08361","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/2106.08361/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":"2106.08361","created_at":"2026-07-05T02:49:57.029221+00:00"},{"alias_kind":"arxiv_version","alias_value":"2106.08361v1","created_at":"2026-07-05T02:49:57.029221+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2106.08361","created_at":"2026-07-05T02:49:57.029221+00:00"},{"alias_kind":"pith_short_12","alias_value":"VK2PEP62OCYW","created_at":"2026-07-05T02:49:57.029221+00:00"},{"alias_kind":"pith_short_16","alias_value":"VK2PEP62OCYWXHMY","created_at":"2026-07-05T02:49:57.029221+00:00"},{"alias_kind":"pith_short_8","alias_value":"VK2PEP62","created_at":"2026-07-05T02:49:57.029221+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2505.02360","citing_title":"Catastrophic Overfitting, Entropy Gap and Participation Ratio: A Noiseless $l^p$ Norm Solution for Fast Adversarial Training","ref_index":13,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/VK2PEP62OCYWXHMY4S5D6VQC2P","json":"https://pith.science/pith/VK2PEP62OCYWXHMY4S5D6VQC2P.json","graph_json":"https://pith.science/api/pith-number/VK2PEP62OCYWXHMY4S5D6VQC2P/graph.json","events_json":"https://pith.science/api/pith-number/VK2PEP62OCYWXHMY4S5D6VQC2P/events.json","paper":"https://pith.science/paper/VK2PEP62"},"agent_actions":{"view_html":"https://pith.science/pith/VK2PEP62OCYWXHMY4S5D6VQC2P","download_json":"https://pith.science/pith/VK2PEP62OCYWXHMY4S5D6VQC2P.json","view_paper":"https://pith.science/paper/VK2PEP62","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2106.08361&json=true","fetch_graph":"https://pith.science/api/pith-number/VK2PEP62OCYWXHMY4S5D6VQC2P/graph.json","fetch_events":"https://pith.science/api/pith-number/VK2PEP62OCYWXHMY4S5D6VQC2P/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/VK2PEP62OCYWXHMY4S5D6VQC2P/action/timestamp_anchor","attest_storage":"https://pith.science/pith/VK2PEP62OCYWXHMY4S5D6VQC2P/action/storage_attestation","attest_author":"https://pith.science/pith/VK2PEP62OCYWXHMY4S5D6VQC2P/action/author_attestation","sign_citation":"https://pith.science/pith/VK2PEP62OCYWXHMY4S5D6VQC2P/action/citation_signature","submit_replication":"https://pith.science/pith/VK2PEP62OCYWXHMY4S5D6VQC2P/action/replication_record"}},"created_at":"2026-07-05T02:49:57.029221+00:00","updated_at":"2026-07-05T02:49:57.029221+00:00"}