{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:FHZMI3HX7J24XFAJY3WDALDZPF","short_pith_number":"pith:FHZMI3HX","schema_version":"1.0","canonical_sha256":"29f2c46cf7fa75cb9409c6ec302c79796570f2cecef4be4afb201077a1d4ef19","source":{"kind":"arxiv","id":"2308.12553","version":1},"attestation_state":"computed","paper":{"title":"Don't blame Dataset Shift! Shortcut Learning due to Gradients and Cross Entropy","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["stat.ML"],"primary_cat":"cs.LG","authors_text":"Aahlad Puli, Lily Zhang, Rajesh Ranganath, Yoav Wald","submitted_at":"2023-08-24T04:39:25Z","abstract_excerpt":"Common explanations for shortcut learning assume that the shortcut improves prediction under the training distribution but not in the test distribution. Thus, models trained via the typical gradient-based optimization of cross-entropy, which we call default-ERM, utilize the shortcut. However, even when the stable feature determines the label in the training distribution and the shortcut does not provide any additional information, like in perception tasks, default-ERM still exhibits shortcut learning. Why are such solutions preferred when the loss for default-ERM can be driven to zero using th"},"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":"2308.12553","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2023-08-24T04:39:25Z","cross_cats_sorted":["stat.ML"],"title_canon_sha256":"c3dad05d2a9e30f3956b3d0351eaa914c009c25c5dfeafe385c4e7fdbbc2ab33","abstract_canon_sha256":"a8bc4a644b0e95aaab1f504e919ce802301cc2e8a2d204c3c0ab55bcbe015a41"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T06:44:14.328268Z","signature_b64":"70iMdKnLkEXpaKw5n0mZt6X4l2ILa3pEPVRCApQyXUqwshTqjVLhLxwG/tiyr766xuv0pvZNpgn6Qgvk6qU3Cw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"29f2c46cf7fa75cb9409c6ec302c79796570f2cecef4be4afb201077a1d4ef19","last_reissued_at":"2026-07-05T06:44:14.327857Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T06:44:14.327857Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Don't blame Dataset Shift! Shortcut Learning due to Gradients and Cross Entropy","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["stat.ML"],"primary_cat":"cs.LG","authors_text":"Aahlad Puli, Lily Zhang, Rajesh Ranganath, Yoav Wald","submitted_at":"2023-08-24T04:39:25Z","abstract_excerpt":"Common explanations for shortcut learning assume that the shortcut improves prediction under the training distribution but not in the test distribution. Thus, models trained via the typical gradient-based optimization of cross-entropy, which we call default-ERM, utilize the shortcut. However, even when the stable feature determines the label in the training distribution and the shortcut does not provide any additional information, like in perception tasks, default-ERM still exhibits shortcut learning. Why are such solutions preferred when the loss for default-ERM can be driven to zero using th"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2308.12553","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/2308.12553/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":"2308.12553","created_at":"2026-07-05T06:44:14.327928+00:00"},{"alias_kind":"arxiv_version","alias_value":"2308.12553v1","created_at":"2026-07-05T06:44:14.327928+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2308.12553","created_at":"2026-07-05T06:44:14.327928+00:00"},{"alias_kind":"pith_short_12","alias_value":"FHZMI3HX7J24","created_at":"2026-07-05T06:44:14.327928+00:00"},{"alias_kind":"pith_short_16","alias_value":"FHZMI3HX7J24XFAJ","created_at":"2026-07-05T06:44:14.327928+00:00"},{"alias_kind":"pith_short_8","alias_value":"FHZMI3HX","created_at":"2026-07-05T06:44:14.327928+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/FHZMI3HX7J24XFAJY3WDALDZPF","json":"https://pith.science/pith/FHZMI3HX7J24XFAJY3WDALDZPF.json","graph_json":"https://pith.science/api/pith-number/FHZMI3HX7J24XFAJY3WDALDZPF/graph.json","events_json":"https://pith.science/api/pith-number/FHZMI3HX7J24XFAJY3WDALDZPF/events.json","paper":"https://pith.science/paper/FHZMI3HX"},"agent_actions":{"view_html":"https://pith.science/pith/FHZMI3HX7J24XFAJY3WDALDZPF","download_json":"https://pith.science/pith/FHZMI3HX7J24XFAJY3WDALDZPF.json","view_paper":"https://pith.science/paper/FHZMI3HX","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2308.12553&json=true","fetch_graph":"https://pith.science/api/pith-number/FHZMI3HX7J24XFAJY3WDALDZPF/graph.json","fetch_events":"https://pith.science/api/pith-number/FHZMI3HX7J24XFAJY3WDALDZPF/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/FHZMI3HX7J24XFAJY3WDALDZPF/action/timestamp_anchor","attest_storage":"https://pith.science/pith/FHZMI3HX7J24XFAJY3WDALDZPF/action/storage_attestation","attest_author":"https://pith.science/pith/FHZMI3HX7J24XFAJY3WDALDZPF/action/author_attestation","sign_citation":"https://pith.science/pith/FHZMI3HX7J24XFAJY3WDALDZPF/action/citation_signature","submit_replication":"https://pith.science/pith/FHZMI3HX7J24XFAJY3WDALDZPF/action/replication_record"}},"created_at":"2026-07-05T06:44:14.327928+00:00","updated_at":"2026-07-05T06:44:14.327928+00:00"}