{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:JDTXCDLOZPDTRXUCL35INFFWYO","short_pith_number":"pith:JDTXCDLO","schema_version":"1.0","canonical_sha256":"48e7710d6ecbc738de825efa8694b6c3b7a9a54ccfa5b75b5275367235b7ec15","source":{"kind":"arxiv","id":"2410.14375","version":3},"attestation_state":"computed","paper":{"title":"Causal Fine-Tuning under Latent Confounded Shift","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.LG","authors_text":"Haoxuan Li, Jialin Yu, Junchi Yu, Mengyue Yang, Nevin L. Zhang, Philip Torr, Ricardo Silva, Yulan He, Yuxiang Zhou","submitted_at":"2024-10-18T11:06:23Z","abstract_excerpt":"Adapting to latent confounded shift remains a core challenge in modern AI. This setting is driven by hidden variables that induce spurious correlations between inputs and outputs during training, leading models to rely on non-causal shortcuts. For example, a model may learn to treat metadata (e.g., data source like \"Amazon\") as a proxy for positive sentiment, causing failure when the source becomes predominantly negative during deployment. To address this latent confounded shift, we introduce Causal Fine-Tuning(CFT). Using a structural causal model as an inductive bias, we derive sufficient id"},"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":"2410.14375","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-10-18T11:06:23Z","cross_cats_sorted":["cs.CL"],"title_canon_sha256":"3c41d8d0a8eac7fddd07590f4a6253e1e85ae0c8656d61e8d729c3f5a2c1e619","abstract_canon_sha256":"1257c864643d337acd0fa4cc113f10ff23cf1be4272a57b7d162dadab62f3118"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-18T03:09:46.439840Z","signature_b64":"9PhM6lPjiPuR0bGQx2vAjYsO24Y+Bh4xPnvP69k2BAS3OFvyr/j+VnY2sU/NfR1//L+fbKEScuxUiXNOqmU4Bg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"48e7710d6ecbc738de825efa8694b6c3b7a9a54ccfa5b75b5275367235b7ec15","last_reissued_at":"2026-05-18T03:09:46.439164Z","signature_status":"signed_v1","first_computed_at":"2026-05-18T03:09:46.439164Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Causal Fine-Tuning under Latent Confounded Shift","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.LG","authors_text":"Haoxuan Li, Jialin Yu, Junchi Yu, Mengyue Yang, Nevin L. Zhang, Philip Torr, Ricardo Silva, Yulan He, Yuxiang Zhou","submitted_at":"2024-10-18T11:06:23Z","abstract_excerpt":"Adapting to latent confounded shift remains a core challenge in modern AI. This setting is driven by hidden variables that induce spurious correlations between inputs and outputs during training, leading models to rely on non-causal shortcuts. For example, a model may learn to treat metadata (e.g., data source like \"Amazon\") as a proxy for positive sentiment, causing failure when the source becomes predominantly negative during deployment. To address this latent confounded shift, we introduce Causal Fine-Tuning(CFT). Using a structural causal model as an inductive bias, we derive sufficient id"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2410.14375","kind":"arxiv","version":3},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"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":"2410.14375","created_at":"2026-05-18T03:09:46.439266+00:00"},{"alias_kind":"arxiv_version","alias_value":"2410.14375v3","created_at":"2026-05-18T03:09:46.439266+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2410.14375","created_at":"2026-05-18T03:09:46.439266+00:00"},{"alias_kind":"pith_short_12","alias_value":"JDTXCDLOZPDT","created_at":"2026-05-18T12:33:37.589309+00:00"},{"alias_kind":"pith_short_16","alias_value":"JDTXCDLOZPDTRXUC","created_at":"2026-05-18T12:33:37.589309+00:00"},{"alias_kind":"pith_short_8","alias_value":"JDTXCDLO","created_at":"2026-05-18T12:33:37.589309+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/JDTXCDLOZPDTRXUCL35INFFWYO","json":"https://pith.science/pith/JDTXCDLOZPDTRXUCL35INFFWYO.json","graph_json":"https://pith.science/api/pith-number/JDTXCDLOZPDTRXUCL35INFFWYO/graph.json","events_json":"https://pith.science/api/pith-number/JDTXCDLOZPDTRXUCL35INFFWYO/events.json","paper":"https://pith.science/paper/JDTXCDLO"},"agent_actions":{"view_html":"https://pith.science/pith/JDTXCDLOZPDTRXUCL35INFFWYO","download_json":"https://pith.science/pith/JDTXCDLOZPDTRXUCL35INFFWYO.json","view_paper":"https://pith.science/paper/JDTXCDLO","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2410.14375&json=true","fetch_graph":"https://pith.science/api/pith-number/JDTXCDLOZPDTRXUCL35INFFWYO/graph.json","fetch_events":"https://pith.science/api/pith-number/JDTXCDLOZPDTRXUCL35INFFWYO/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/JDTXCDLOZPDTRXUCL35INFFWYO/action/timestamp_anchor","attest_storage":"https://pith.science/pith/JDTXCDLOZPDTRXUCL35INFFWYO/action/storage_attestation","attest_author":"https://pith.science/pith/JDTXCDLOZPDTRXUCL35INFFWYO/action/author_attestation","sign_citation":"https://pith.science/pith/JDTXCDLOZPDTRXUCL35INFFWYO/action/citation_signature","submit_replication":"https://pith.science/pith/JDTXCDLOZPDTRXUCL35INFFWYO/action/replication_record"}},"created_at":"2026-05-18T03:09:46.439266+00:00","updated_at":"2026-05-18T03:09:46.439266+00:00"}