{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2020:VMKXUTU2IGSBRG674Y5NAKWNO6","short_pith_number":"pith:VMKXUTU2","schema_version":"1.0","canonical_sha256":"ab157a4e9a41a4189bdfe63ad02acd77aa28aab66b032b0a2d802745444d30cb","source":{"kind":"arxiv","id":"2010.04819","version":4},"attestation_state":"computed","paper":{"title":"How Does Mixup Help With Robustness and Generalization?","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["stat.ML"],"primary_cat":"cs.LG","authors_text":"Amirata Ghorbani, James Zou, Kenji Kawaguchi, Linjun Zhang, Zhun Deng","submitted_at":"2020-10-09T21:38:14Z","abstract_excerpt":"Mixup is a popular data augmentation technique based on taking convex combinations of pairs of examples and their labels. This simple technique has been shown to substantially improve both the robustness and the generalization of the trained model. However, it is not well-understood why such improvement occurs. In this paper, we provide theoretical analysis to demonstrate how using Mixup in training helps model robustness and generalization. For robustness, we show that minimizing the Mixup loss corresponds to approximately minimizing an upper bound of the adversarial loss. This explains why m"},"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":"2010.04819","kind":"arxiv","version":4},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2020-10-09T21:38:14Z","cross_cats_sorted":["stat.ML"],"title_canon_sha256":"7e0fab3444e11d39b568826df146f71ed006e5e48f2c13ac1b2398c2eb54a5b7","abstract_canon_sha256":"e134b07acbaf077927fab8792ff3a9b2c1c4af0768244f78bdcfebfb5c0128ad"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T02:24:11.361287Z","signature_b64":"uNaKoCF+tKiEN5Cs2hXLOEHrd8GRRheaTAM/j3npV56nlv1lMQCIx5EuoU5ZPEEohRjpGC/PXCIZop4kXg/4DQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"ab157a4e9a41a4189bdfe63ad02acd77aa28aab66b032b0a2d802745444d30cb","last_reissued_at":"2026-07-05T02:24:11.360834Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T02:24:11.360834Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"How Does Mixup Help With Robustness and Generalization?","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["stat.ML"],"primary_cat":"cs.LG","authors_text":"Amirata Ghorbani, James Zou, Kenji Kawaguchi, Linjun Zhang, Zhun Deng","submitted_at":"2020-10-09T21:38:14Z","abstract_excerpt":"Mixup is a popular data augmentation technique based on taking convex combinations of pairs of examples and their labels. This simple technique has been shown to substantially improve both the robustness and the generalization of the trained model. However, it is not well-understood why such improvement occurs. In this paper, we provide theoretical analysis to demonstrate how using Mixup in training helps model robustness and generalization. For robustness, we show that minimizing the Mixup loss corresponds to approximately minimizing an upper bound of the adversarial loss. This explains why m"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2010.04819","kind":"arxiv","version":4},"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/2010.04819/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":"2010.04819","created_at":"2026-07-05T02:24:11.360892+00:00"},{"alias_kind":"arxiv_version","alias_value":"2010.04819v4","created_at":"2026-07-05T02:24:11.360892+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2010.04819","created_at":"2026-07-05T02:24:11.360892+00:00"},{"alias_kind":"pith_short_12","alias_value":"VMKXUTU2IGSB","created_at":"2026-07-05T02:24:11.360892+00:00"},{"alias_kind":"pith_short_16","alias_value":"VMKXUTU2IGSBRG67","created_at":"2026-07-05T02:24:11.360892+00:00"},{"alias_kind":"pith_short_8","alias_value":"VMKXUTU2","created_at":"2026-07-05T02:24:11.360892+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":4,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.15416","citing_title":"Margin-Adaptive Confidence Ranking for Reliable LLM Judgement","ref_index":78,"is_internal_anchor":false},{"citing_arxiv_id":"2605.09716","citing_title":"Medical Model Synthesis Architectures: A Case Study","ref_index":182,"is_internal_anchor":false},{"citing_arxiv_id":"2605.10290","citing_title":"Characterizing the Generalization Error of Random Feature Regression with Arbitrary Data-Augmentation","ref_index":5,"is_internal_anchor":false},{"citing_arxiv_id":"2604.19941","citing_title":"CrackForward: Context-Aware Severity Stage Crack Synthesis for Data Augmentation","ref_index":16,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/VMKXUTU2IGSBRG674Y5NAKWNO6","json":"https://pith.science/pith/VMKXUTU2IGSBRG674Y5NAKWNO6.json","graph_json":"https://pith.science/api/pith-number/VMKXUTU2IGSBRG674Y5NAKWNO6/graph.json","events_json":"https://pith.science/api/pith-number/VMKXUTU2IGSBRG674Y5NAKWNO6/events.json","paper":"https://pith.science/paper/VMKXUTU2"},"agent_actions":{"view_html":"https://pith.science/pith/VMKXUTU2IGSBRG674Y5NAKWNO6","download_json":"https://pith.science/pith/VMKXUTU2IGSBRG674Y5NAKWNO6.json","view_paper":"https://pith.science/paper/VMKXUTU2","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2010.04819&json=true","fetch_graph":"https://pith.science/api/pith-number/VMKXUTU2IGSBRG674Y5NAKWNO6/graph.json","fetch_events":"https://pith.science/api/pith-number/VMKXUTU2IGSBRG674Y5NAKWNO6/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/VMKXUTU2IGSBRG674Y5NAKWNO6/action/timestamp_anchor","attest_storage":"https://pith.science/pith/VMKXUTU2IGSBRG674Y5NAKWNO6/action/storage_attestation","attest_author":"https://pith.science/pith/VMKXUTU2IGSBRG674Y5NAKWNO6/action/author_attestation","sign_citation":"https://pith.science/pith/VMKXUTU2IGSBRG674Y5NAKWNO6/action/citation_signature","submit_replication":"https://pith.science/pith/VMKXUTU2IGSBRG674Y5NAKWNO6/action/replication_record"}},"created_at":"2026-07-05T02:24:11.360892+00:00","updated_at":"2026-07-05T02:24:11.360892+00:00"}