{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:KMGLNZFLMPI6AYMTA6BCVI7U4X","short_pith_number":"pith:KMGLNZFL","schema_version":"1.0","canonical_sha256":"530cb6e4ab63d1e0619307822aa3f4e5c6198462a5b9937fcaba6e2f3485dae7","source":{"kind":"arxiv","id":"2303.01475","version":1},"attestation_state":"computed","paper":{"title":"Over-training with Mixup May Hurt Generalization","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Hongyu Guo, Yongyi Mao, Ziqiao Wang, Zixuan Liu","submitted_at":"2023-03-02T18:37:34Z","abstract_excerpt":"Mixup, which creates synthetic training instances by linearly interpolating random sample pairs, is a simple and yet effective regularization technique to boost the performance of deep models trained with SGD. In this work, we report a previously unobserved phenomenon in Mixup training: on a number of standard datasets, the performance of Mixup-trained models starts to decay after training for a large number of epochs, giving rise to a U-shaped generalization curve. This behavior is further aggravated when the size of original dataset is reduced. To help understand such a behavior of Mixup, we"},"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":"2303.01475","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2023-03-02T18:37:34Z","cross_cats_sorted":[],"title_canon_sha256":"775b61caf3bbfc92ff40ffdb27c8340dc01688299ab9818eec8bed57e9c48cb7","abstract_canon_sha256":"bbb1b40c94293c1258e8a5eba58cbf404c6b58fccade2dae629aa8970d77b900"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T05:47:31.437011Z","signature_b64":"rjeLOv7pNp1vOtWXpFNyy53BzYL4Rb/5E5cj8e/vFSd/wZhEoe/IJPslNSGsjLe2k+yEEQCV4GtI9up4z/OEAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"530cb6e4ab63d1e0619307822aa3f4e5c6198462a5b9937fcaba6e2f3485dae7","last_reissued_at":"2026-07-05T05:47:31.436597Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T05:47:31.436597Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Over-training with Mixup May Hurt Generalization","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Hongyu Guo, Yongyi Mao, Ziqiao Wang, Zixuan Liu","submitted_at":"2023-03-02T18:37:34Z","abstract_excerpt":"Mixup, which creates synthetic training instances by linearly interpolating random sample pairs, is a simple and yet effective regularization technique to boost the performance of deep models trained with SGD. In this work, we report a previously unobserved phenomenon in Mixup training: on a number of standard datasets, the performance of Mixup-trained models starts to decay after training for a large number of epochs, giving rise to a U-shaped generalization curve. This behavior is further aggravated when the size of original dataset is reduced. To help understand such a behavior of Mixup, we"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2303.01475","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/2303.01475/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":"2303.01475","created_at":"2026-07-05T05:47:31.436657+00:00"},{"alias_kind":"arxiv_version","alias_value":"2303.01475v1","created_at":"2026-07-05T05:47:31.436657+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2303.01475","created_at":"2026-07-05T05:47:31.436657+00:00"},{"alias_kind":"pith_short_12","alias_value":"KMGLNZFLMPI6","created_at":"2026-07-05T05:47:31.436657+00:00"},{"alias_kind":"pith_short_16","alias_value":"KMGLNZFLMPI6AYMT","created_at":"2026-07-05T05:47:31.436657+00:00"},{"alias_kind":"pith_short_8","alias_value":"KMGLNZFL","created_at":"2026-07-05T05:47:31.436657+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2502.00412","citing_title":"TROI: Cross-Subject Pretraining with Sparse Voxel Selection for Enhanced fMRI Visual Decoding","ref_index":16,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/KMGLNZFLMPI6AYMTA6BCVI7U4X","json":"https://pith.science/pith/KMGLNZFLMPI6AYMTA6BCVI7U4X.json","graph_json":"https://pith.science/api/pith-number/KMGLNZFLMPI6AYMTA6BCVI7U4X/graph.json","events_json":"https://pith.science/api/pith-number/KMGLNZFLMPI6AYMTA6BCVI7U4X/events.json","paper":"https://pith.science/paper/KMGLNZFL"},"agent_actions":{"view_html":"https://pith.science/pith/KMGLNZFLMPI6AYMTA6BCVI7U4X","download_json":"https://pith.science/pith/KMGLNZFLMPI6AYMTA6BCVI7U4X.json","view_paper":"https://pith.science/paper/KMGLNZFL","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2303.01475&json=true","fetch_graph":"https://pith.science/api/pith-number/KMGLNZFLMPI6AYMTA6BCVI7U4X/graph.json","fetch_events":"https://pith.science/api/pith-number/KMGLNZFLMPI6AYMTA6BCVI7U4X/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/KMGLNZFLMPI6AYMTA6BCVI7U4X/action/timestamp_anchor","attest_storage":"https://pith.science/pith/KMGLNZFLMPI6AYMTA6BCVI7U4X/action/storage_attestation","attest_author":"https://pith.science/pith/KMGLNZFLMPI6AYMTA6BCVI7U4X/action/author_attestation","sign_citation":"https://pith.science/pith/KMGLNZFLMPI6AYMTA6BCVI7U4X/action/citation_signature","submit_replication":"https://pith.science/pith/KMGLNZFLMPI6AYMTA6BCVI7U4X/action/replication_record"}},"created_at":"2026-07-05T05:47:31.436657+00:00","updated_at":"2026-07-05T05:47:31.436657+00:00"}