{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:NJVLHYUG5MNYLSYTX2MVVBFS2M","short_pith_number":"pith:NJVLHYUG","schema_version":"1.0","canonical_sha256":"6a6ab3e286eb1b85cb13be995a84b2d30cf19f44fe0788395cd330d637a261b4","source":{"kind":"arxiv","id":"2109.07048","version":2},"attestation_state":"computed","paper":{"title":"ARCH: Efficient Adversarial Regularized Training with Caching","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Chen Liang, Haoming Jiang, Jianfeng Gao, Pengcheng He, Simiao Zuo, Tuo Zhao, Weizhu Chen, XiaoDong Liu","submitted_at":"2021-09-15T02:05:37Z","abstract_excerpt":"Adversarial regularization can improve model generalization in many natural language processing tasks. However, conventional approaches are computationally expensive since they need to generate a perturbation for each sample in each epoch. We propose a new adversarial regularization method ARCH (adversarial regularization with caching), where perturbations are generated and cached once every several epochs. As caching all the perturbations imposes memory usage concerns, we adopt a K-nearest neighbors-based strategy to tackle this issue. The strategy only requires caching a small amount of pert"},"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":"2109.07048","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2021-09-15T02:05:37Z","cross_cats_sorted":[],"title_canon_sha256":"e146d6b2768c9b72165d768bcdd0298a465fa2693ce5aa8419514676f53a0009","abstract_canon_sha256":"660844130d5a89a078430875d96cd175cadce25146d3a3964dc94dadba1dc2ce"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T04:16:30.841063Z","signature_b64":"rSKJR8WXpigT/48rIXCIWIzVSJFLiYpMG2P883hrLAk6ewxUfAAjqj/E/+OMRsnxH4SXqb0Tq3/s6yI2e78rAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"6a6ab3e286eb1b85cb13be995a84b2d30cf19f44fe0788395cd330d637a261b4","last_reissued_at":"2026-07-05T04:16:30.840649Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T04:16:30.840649Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"ARCH: Efficient Adversarial Regularized Training with Caching","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Chen Liang, Haoming Jiang, Jianfeng Gao, Pengcheng He, Simiao Zuo, Tuo Zhao, Weizhu Chen, XiaoDong Liu","submitted_at":"2021-09-15T02:05:37Z","abstract_excerpt":"Adversarial regularization can improve model generalization in many natural language processing tasks. However, conventional approaches are computationally expensive since they need to generate a perturbation for each sample in each epoch. We propose a new adversarial regularization method ARCH (adversarial regularization with caching), where perturbations are generated and cached once every several epochs. As caching all the perturbations imposes memory usage concerns, we adopt a K-nearest neighbors-based strategy to tackle this issue. The strategy only requires caching a small amount of pert"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2109.07048","kind":"arxiv","version":2},"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/2109.07048/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":"2109.07048","created_at":"2026-07-05T04:16:30.840700+00:00"},{"alias_kind":"arxiv_version","alias_value":"2109.07048v2","created_at":"2026-07-05T04:16:30.840700+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2109.07048","created_at":"2026-07-05T04:16:30.840700+00:00"},{"alias_kind":"pith_short_12","alias_value":"NJVLHYUG5MNY","created_at":"2026-07-05T04:16:30.840700+00:00"},{"alias_kind":"pith_short_16","alias_value":"NJVLHYUG5MNYLSYT","created_at":"2026-07-05T04:16:30.840700+00:00"},{"alias_kind":"pith_short_8","alias_value":"NJVLHYUG","created_at":"2026-07-05T04:16:30.840700+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/NJVLHYUG5MNYLSYTX2MVVBFS2M","json":"https://pith.science/pith/NJVLHYUG5MNYLSYTX2MVVBFS2M.json","graph_json":"https://pith.science/api/pith-number/NJVLHYUG5MNYLSYTX2MVVBFS2M/graph.json","events_json":"https://pith.science/api/pith-number/NJVLHYUG5MNYLSYTX2MVVBFS2M/events.json","paper":"https://pith.science/paper/NJVLHYUG"},"agent_actions":{"view_html":"https://pith.science/pith/NJVLHYUG5MNYLSYTX2MVVBFS2M","download_json":"https://pith.science/pith/NJVLHYUG5MNYLSYTX2MVVBFS2M.json","view_paper":"https://pith.science/paper/NJVLHYUG","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2109.07048&json=true","fetch_graph":"https://pith.science/api/pith-number/NJVLHYUG5MNYLSYTX2MVVBFS2M/graph.json","fetch_events":"https://pith.science/api/pith-number/NJVLHYUG5MNYLSYTX2MVVBFS2M/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/NJVLHYUG5MNYLSYTX2MVVBFS2M/action/timestamp_anchor","attest_storage":"https://pith.science/pith/NJVLHYUG5MNYLSYTX2MVVBFS2M/action/storage_attestation","attest_author":"https://pith.science/pith/NJVLHYUG5MNYLSYTX2MVVBFS2M/action/author_attestation","sign_citation":"https://pith.science/pith/NJVLHYUG5MNYLSYTX2MVVBFS2M/action/citation_signature","submit_replication":"https://pith.science/pith/NJVLHYUG5MNYLSYTX2MVVBFS2M/action/replication_record"}},"created_at":"2026-07-05T04:16:30.840700+00:00","updated_at":"2026-07-05T04:16:30.840700+00:00"}