{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:AJ3JHIBOC2KY2NL64DP6SVBZUM","short_pith_number":"pith:AJ3JHIBO","schema_version":"1.0","canonical_sha256":"027693a02e16958d357ee0dfe95439a32b576130c62d9a5bc5c50803449e2b35","source":{"kind":"arxiv","id":"2407.13957","version":2},"attestation_state":"computed","paper":{"title":"The Group Robustness is in the Details: Revisiting Finetuning under Spurious Correlations","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Abhishek Kumar, John C. Hill, Tyler LaBonte, Vidya Muthukumar, Xinchen Zhang","submitted_at":"2024-07-19T00:34:03Z","abstract_excerpt":"Modern machine learning models are prone to over-reliance on spurious correlations, which can often lead to poor performance on minority groups. In this paper, we identify surprising and nuanced behavior of finetuned models on worst-group accuracy via comprehensive experiments on four well-established benchmarks across vision and language tasks. We first show that the commonly used class-balancing techniques of mini-batch upsampling and loss upweighting can induce a decrease in worst-group accuracy (WGA) with training epochs, leading to performance no better than without class-balancing. While"},"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":"2407.13957","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-07-19T00:34:03Z","cross_cats_sorted":[],"title_canon_sha256":"ee6e087fe53a36046116e1aaa01f82f48305c0fbeb8f1e0bc627a36cc7d55112","abstract_canon_sha256":"8feb35fab3404f7e95b5ca48c0d9b3b6ad3ffe701c5f11e41083cc37660ef1b0"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:27:03.309235Z","signature_b64":"kRTDSUJIUjGzY3gHL57CzeTF2VAqyeCu7eOzACZ8X/nQBhMilaxoRkytpdUCxhm4zNgms9Nhe21lZD+bpa9TAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"027693a02e16958d357ee0dfe95439a32b576130c62d9a5bc5c50803449e2b35","last_reissued_at":"2026-07-05T09:27:03.308696Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:27:03.308696Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"The Group Robustness is in the Details: Revisiting Finetuning under Spurious Correlations","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Abhishek Kumar, John C. Hill, Tyler LaBonte, Vidya Muthukumar, Xinchen Zhang","submitted_at":"2024-07-19T00:34:03Z","abstract_excerpt":"Modern machine learning models are prone to over-reliance on spurious correlations, which can often lead to poor performance on minority groups. In this paper, we identify surprising and nuanced behavior of finetuned models on worst-group accuracy via comprehensive experiments on four well-established benchmarks across vision and language tasks. We first show that the commonly used class-balancing techniques of mini-batch upsampling and loss upweighting can induce a decrease in worst-group accuracy (WGA) with training epochs, leading to performance no better than without class-balancing. While"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2407.13957","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/2407.13957/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":"2407.13957","created_at":"2026-07-05T09:27:03.308747+00:00"},{"alias_kind":"arxiv_version","alias_value":"2407.13957v2","created_at":"2026-07-05T09:27:03.308747+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2407.13957","created_at":"2026-07-05T09:27:03.308747+00:00"},{"alias_kind":"pith_short_12","alias_value":"AJ3JHIBOC2KY","created_at":"2026-07-05T09:27:03.308747+00:00"},{"alias_kind":"pith_short_16","alias_value":"AJ3JHIBOC2KY2NL6","created_at":"2026-07-05T09:27:03.308747+00:00"},{"alias_kind":"pith_short_8","alias_value":"AJ3JHIBO","created_at":"2026-07-05T09:27:03.308747+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2506.11347","citing_title":"Improving Group Robustness on Spurious Correlation via Evidential Alignment","ref_index":20,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/AJ3JHIBOC2KY2NL64DP6SVBZUM","json":"https://pith.science/pith/AJ3JHIBOC2KY2NL64DP6SVBZUM.json","graph_json":"https://pith.science/api/pith-number/AJ3JHIBOC2KY2NL64DP6SVBZUM/graph.json","events_json":"https://pith.science/api/pith-number/AJ3JHIBOC2KY2NL64DP6SVBZUM/events.json","paper":"https://pith.science/paper/AJ3JHIBO"},"agent_actions":{"view_html":"https://pith.science/pith/AJ3JHIBOC2KY2NL64DP6SVBZUM","download_json":"https://pith.science/pith/AJ3JHIBOC2KY2NL64DP6SVBZUM.json","view_paper":"https://pith.science/paper/AJ3JHIBO","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2407.13957&json=true","fetch_graph":"https://pith.science/api/pith-number/AJ3JHIBOC2KY2NL64DP6SVBZUM/graph.json","fetch_events":"https://pith.science/api/pith-number/AJ3JHIBOC2KY2NL64DP6SVBZUM/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/AJ3JHIBOC2KY2NL64DP6SVBZUM/action/timestamp_anchor","attest_storage":"https://pith.science/pith/AJ3JHIBOC2KY2NL64DP6SVBZUM/action/storage_attestation","attest_author":"https://pith.science/pith/AJ3JHIBOC2KY2NL64DP6SVBZUM/action/author_attestation","sign_citation":"https://pith.science/pith/AJ3JHIBOC2KY2NL64DP6SVBZUM/action/citation_signature","submit_replication":"https://pith.science/pith/AJ3JHIBOC2KY2NL64DP6SVBZUM/action/replication_record"}},"created_at":"2026-07-05T09:27:03.308747+00:00","updated_at":"2026-07-05T09:27:03.308747+00:00"}