{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:L6WAR5OROXHJP64L43NPC4XLIJ","short_pith_number":"pith:L6WAR5OR","schema_version":"1.0","canonical_sha256":"5fac08f5d175ce97fb8be6daf172eb4256dbd33bfbb1c4215c3248dad5c5f410","source":{"kind":"arxiv","id":"2110.14503","version":2},"attestation_state":"computed","paper":{"title":"Simple data balancing achieves competitive worst-group-accuracy","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.CR"],"primary_cat":"cs.LG","authors_text":"Badr Youbi Idrissi, David Lopez-Paz, Martin Arjovsky, Mohammad Pezeshki","submitted_at":"2021-10-27T15:15:11Z","abstract_excerpt":"We study the problem of learning classifiers that perform well across (known or unknown) groups of data. After observing that common worst-group-accuracy datasets suffer from substantial imbalances, we set out to compare state-of-the-art methods to simple balancing of classes and groups by either subsampling or reweighting data. Our results show that these data balancing baselines achieve state-of-the-art-accuracy, while being faster to train and requiring no additional hyper-parameters. In addition, we highlight that access to group information is most critical for model selection purposes, a"},"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":"2110.14503","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2021-10-27T15:15:11Z","cross_cats_sorted":["cs.AI","cs.CR"],"title_canon_sha256":"0b9f91564c05dbecf31a3e2e9a5d3a3ca3e2f8ce39499fdb0c704d0e0f2cf828","abstract_canon_sha256":"b6ec584f1fb9dcc7062378457366f72995e344d49e0d5c6a244e5dcd2e6cd1b0"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T03:58:04.607109Z","signature_b64":"9Me67ObHBRsTIYWhCD24ZhHmeE38g3WsFEYG1tXMCj2IvOrOhVW6XWoGW3OIKPCqvNB9IzeoR/dQMlQoh7UnAQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"5fac08f5d175ce97fb8be6daf172eb4256dbd33bfbb1c4215c3248dad5c5f410","last_reissued_at":"2026-07-05T03:58:04.606630Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T03:58:04.606630Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Simple data balancing achieves competitive worst-group-accuracy","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.CR"],"primary_cat":"cs.LG","authors_text":"Badr Youbi Idrissi, David Lopez-Paz, Martin Arjovsky, Mohammad Pezeshki","submitted_at":"2021-10-27T15:15:11Z","abstract_excerpt":"We study the problem of learning classifiers that perform well across (known or unknown) groups of data. After observing that common worst-group-accuracy datasets suffer from substantial imbalances, we set out to compare state-of-the-art methods to simple balancing of classes and groups by either subsampling or reweighting data. Our results show that these data balancing baselines achieve state-of-the-art-accuracy, while being faster to train and requiring no additional hyper-parameters. In addition, we highlight that access to group information is most critical for model selection purposes, a"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2110.14503","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/2110.14503/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":"2110.14503","created_at":"2026-07-05T03:58:04.606690+00:00"},{"alias_kind":"arxiv_version","alias_value":"2110.14503v2","created_at":"2026-07-05T03:58:04.606690+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2110.14503","created_at":"2026-07-05T03:58:04.606690+00:00"},{"alias_kind":"pith_short_12","alias_value":"L6WAR5OROXHJ","created_at":"2026-07-05T03:58:04.606690+00:00"},{"alias_kind":"pith_short_16","alias_value":"L6WAR5OROXHJP64L","created_at":"2026-07-05T03:58:04.606690+00:00"},{"alias_kind":"pith_short_8","alias_value":"L6WAR5OR","created_at":"2026-07-05T03:58:04.606690+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.24161","citing_title":"Dual-Branch Cross-Projection Debiasing through Diffusion-based Disentanglement","ref_index":24,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/L6WAR5OROXHJP64L43NPC4XLIJ","json":"https://pith.science/pith/L6WAR5OROXHJP64L43NPC4XLIJ.json","graph_json":"https://pith.science/api/pith-number/L6WAR5OROXHJP64L43NPC4XLIJ/graph.json","events_json":"https://pith.science/api/pith-number/L6WAR5OROXHJP64L43NPC4XLIJ/events.json","paper":"https://pith.science/paper/L6WAR5OR"},"agent_actions":{"view_html":"https://pith.science/pith/L6WAR5OROXHJP64L43NPC4XLIJ","download_json":"https://pith.science/pith/L6WAR5OROXHJP64L43NPC4XLIJ.json","view_paper":"https://pith.science/paper/L6WAR5OR","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2110.14503&json=true","fetch_graph":"https://pith.science/api/pith-number/L6WAR5OROXHJP64L43NPC4XLIJ/graph.json","fetch_events":"https://pith.science/api/pith-number/L6WAR5OROXHJP64L43NPC4XLIJ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/L6WAR5OROXHJP64L43NPC4XLIJ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/L6WAR5OROXHJP64L43NPC4XLIJ/action/storage_attestation","attest_author":"https://pith.science/pith/L6WAR5OROXHJP64L43NPC4XLIJ/action/author_attestation","sign_citation":"https://pith.science/pith/L6WAR5OROXHJP64L43NPC4XLIJ/action/citation_signature","submit_replication":"https://pith.science/pith/L6WAR5OROXHJP64L43NPC4XLIJ/action/replication_record"}},"created_at":"2026-07-05T03:58:04.606690+00:00","updated_at":"2026-07-05T03:58:04.606690+00:00"}