{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2022:YDB2N36ATVASSLGJ6ONLAQESHX","short_pith_number":"pith:YDB2N36A","schema_version":"1.0","canonical_sha256":"c0c3a6efc09d41292cc9f39ab040923dc5c9634389278ad60a1b7e612697debb","source":{"kind":"arxiv","id":"2206.12739","version":1},"attestation_state":"computed","paper":{"title":"On how to avoid exacerbating spurious correlations when models are overparameterized","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Christos Thrampoulidis, Ke Wang, Tina Behnia","submitted_at":"2022-06-25T21:53:44Z","abstract_excerpt":"Overparameterized models fail to generalize well in the presence of data imbalance even when combined with traditional techniques for mitigating imbalances. This paper focuses on imbalanced classification datasets, in which a small subset of the population -- a minority -- may contain features that correlate spuriously with the class label. For a parametric family of cross-entropy loss modifications and a representative Gaussian mixture model, we derive non-asymptotic generalization bounds on the worst-group error that shed light on the role of different hyper-parameters. Specifically, we prov"},"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":"2206.12739","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2022-06-25T21:53:44Z","cross_cats_sorted":[],"title_canon_sha256":"ce83e6096a0e007db43f7736f2666450f8ef7e6cf6c7744c2109b6ea78816612","abstract_canon_sha256":"bad09095dab49004a2f846883cd9216bd4c0a59dbd51fc7286fe85ccd7b9d363"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T04:34:58.579529Z","signature_b64":"zTd3Kx4L3ssmoDndFKfwCAozSONI0QJf/4OA2kPkqPKpnM5f1+3OnSgQsP7jImBUc/x9IIRR4RAJsMroEJIZCQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"c0c3a6efc09d41292cc9f39ab040923dc5c9634389278ad60a1b7e612697debb","last_reissued_at":"2026-07-05T04:34:58.579193Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T04:34:58.579193Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"On how to avoid exacerbating spurious correlations when models are overparameterized","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Christos Thrampoulidis, Ke Wang, Tina Behnia","submitted_at":"2022-06-25T21:53:44Z","abstract_excerpt":"Overparameterized models fail to generalize well in the presence of data imbalance even when combined with traditional techniques for mitigating imbalances. This paper focuses on imbalanced classification datasets, in which a small subset of the population -- a minority -- may contain features that correlate spuriously with the class label. For a parametric family of cross-entropy loss modifications and a representative Gaussian mixture model, we derive non-asymptotic generalization bounds on the worst-group error that shed light on the role of different hyper-parameters. Specifically, we prov"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2206.12739","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/2206.12739/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":"2206.12739","created_at":"2026-07-05T04:34:58.579245+00:00"},{"alias_kind":"arxiv_version","alias_value":"2206.12739v1","created_at":"2026-07-05T04:34:58.579245+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2206.12739","created_at":"2026-07-05T04:34:58.579245+00:00"},{"alias_kind":"pith_short_12","alias_value":"YDB2N36ATVAS","created_at":"2026-07-05T04:34:58.579245+00:00"},{"alias_kind":"pith_short_16","alias_value":"YDB2N36ATVASSLGJ","created_at":"2026-07-05T04:34:58.579245+00:00"},{"alias_kind":"pith_short_8","alias_value":"YDB2N36A","created_at":"2026-07-05T04:34:58.579245+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2506.20025","citing_title":"Thumb on the Scale: Optimal Loss Weighting in Last Layer Retraining","ref_index":18,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/YDB2N36ATVASSLGJ6ONLAQESHX","json":"https://pith.science/pith/YDB2N36ATVASSLGJ6ONLAQESHX.json","graph_json":"https://pith.science/api/pith-number/YDB2N36ATVASSLGJ6ONLAQESHX/graph.json","events_json":"https://pith.science/api/pith-number/YDB2N36ATVASSLGJ6ONLAQESHX/events.json","paper":"https://pith.science/paper/YDB2N36A"},"agent_actions":{"view_html":"https://pith.science/pith/YDB2N36ATVASSLGJ6ONLAQESHX","download_json":"https://pith.science/pith/YDB2N36ATVASSLGJ6ONLAQESHX.json","view_paper":"https://pith.science/paper/YDB2N36A","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2206.12739&json=true","fetch_graph":"https://pith.science/api/pith-number/YDB2N36ATVASSLGJ6ONLAQESHX/graph.json","fetch_events":"https://pith.science/api/pith-number/YDB2N36ATVASSLGJ6ONLAQESHX/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/YDB2N36ATVASSLGJ6ONLAQESHX/action/timestamp_anchor","attest_storage":"https://pith.science/pith/YDB2N36ATVASSLGJ6ONLAQESHX/action/storage_attestation","attest_author":"https://pith.science/pith/YDB2N36ATVASSLGJ6ONLAQESHX/action/author_attestation","sign_citation":"https://pith.science/pith/YDB2N36ATVASSLGJ6ONLAQESHX/action/citation_signature","submit_replication":"https://pith.science/pith/YDB2N36ATVASSLGJ6ONLAQESHX/action/replication_record"}},"created_at":"2026-07-05T04:34:58.579245+00:00","updated_at":"2026-07-05T04:34:58.579245+00:00"}