{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2018:WLAIOSUOK25TRUASQSELYGSOL3","short_pith_number":"pith:WLAIOSUO","schema_version":"1.0","canonical_sha256":"b2c0874a8e56bb38d0128488bc1a4e5ef5f3496a9c7cb26ab6f19adad012e2bc","source":{"kind":"arxiv","id":"1809.04198","version":1},"attestation_state":"computed","paper":{"title":"Optimization with Non-Differentiable Constraints with Applications to Fairness, Recall, Churn, and Other Goals","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.GT","math.OC","stat.ML"],"primary_cat":"cs.LG","authors_text":"Andrew Cotter, Heinrich Jiang, Karthik Sridharan, Maya Gupta, Serena Wang, Seungil You, Taman Narayan","submitted_at":"2018-09-11T23:41:47Z","abstract_excerpt":"We show that many machine learning goals, such as improved fairness metrics, can be expressed as constraints on the model's predictions, which we call rate constraints. We study the problem of training non-convex models subject to these rate constraints (or any non-convex and non-differentiable constraints). In the non-convex setting, the standard approach of Lagrange multipliers may fail. Furthermore, if the constraints are non-differentiable, then one cannot optimize the Lagrangian with gradient-based methods. To solve these issues, we introduce the proxy-Lagrangian formulation. This new for"},"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":"1809.04198","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2018-09-11T23:41:47Z","cross_cats_sorted":["cs.AI","cs.GT","math.OC","stat.ML"],"title_canon_sha256":"a8e0abe90e78a2149be655f44ff927d3330d6248ca8aabc433e483c0d857c0e0","abstract_canon_sha256":"750e5edac31a375262482a67ecf491bc724aa6db6e859545e4261ee39dbab644"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-18T00:05:55.215503Z","signature_b64":"F776inRNWkFJOFRN3uXRn7welvLGy8iQUVw+Pue5Afh7QIIUdJOpHpoLFuSy8ZxCDpzNv4XHkVU8JHeUsOi2DA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"b2c0874a8e56bb38d0128488bc1a4e5ef5f3496a9c7cb26ab6f19adad012e2bc","last_reissued_at":"2026-05-18T00:05:55.215042Z","signature_status":"signed_v1","first_computed_at":"2026-05-18T00:05:55.215042Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Optimization with Non-Differentiable Constraints with Applications to Fairness, Recall, Churn, and Other Goals","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.GT","math.OC","stat.ML"],"primary_cat":"cs.LG","authors_text":"Andrew Cotter, Heinrich Jiang, Karthik Sridharan, Maya Gupta, Serena Wang, Seungil You, Taman Narayan","submitted_at":"2018-09-11T23:41:47Z","abstract_excerpt":"We show that many machine learning goals, such as improved fairness metrics, can be expressed as constraints on the model's predictions, which we call rate constraints. We study the problem of training non-convex models subject to these rate constraints (or any non-convex and non-differentiable constraints). In the non-convex setting, the standard approach of Lagrange multipliers may fail. Furthermore, if the constraints are non-differentiable, then one cannot optimize the Lagrangian with gradient-based methods. To solve these issues, we introduce the proxy-Lagrangian formulation. This new for"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1809.04198","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":""},"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":"1809.04198","created_at":"2026-05-18T00:05:55.215113+00:00"},{"alias_kind":"arxiv_version","alias_value":"1809.04198v1","created_at":"2026-05-18T00:05:55.215113+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1809.04198","created_at":"2026-05-18T00:05:55.215113+00:00"},{"alias_kind":"pith_short_12","alias_value":"WLAIOSUOK25T","created_at":"2026-05-18T12:33:01.666342+00:00"},{"alias_kind":"pith_short_16","alias_value":"WLAIOSUOK25TRUAS","created_at":"2026-05-18T12:33:01.666342+00:00"},{"alias_kind":"pith_short_8","alias_value":"WLAIOSUO","created_at":"2026-05-18T12:33:01.666342+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/WLAIOSUOK25TRUASQSELYGSOL3","json":"https://pith.science/pith/WLAIOSUOK25TRUASQSELYGSOL3.json","graph_json":"https://pith.science/api/pith-number/WLAIOSUOK25TRUASQSELYGSOL3/graph.json","events_json":"https://pith.science/api/pith-number/WLAIOSUOK25TRUASQSELYGSOL3/events.json","paper":"https://pith.science/paper/WLAIOSUO"},"agent_actions":{"view_html":"https://pith.science/pith/WLAIOSUOK25TRUASQSELYGSOL3","download_json":"https://pith.science/pith/WLAIOSUOK25TRUASQSELYGSOL3.json","view_paper":"https://pith.science/paper/WLAIOSUO","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=1809.04198&json=true","fetch_graph":"https://pith.science/api/pith-number/WLAIOSUOK25TRUASQSELYGSOL3/graph.json","fetch_events":"https://pith.science/api/pith-number/WLAIOSUOK25TRUASQSELYGSOL3/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/WLAIOSUOK25TRUASQSELYGSOL3/action/timestamp_anchor","attest_storage":"https://pith.science/pith/WLAIOSUOK25TRUASQSELYGSOL3/action/storage_attestation","attest_author":"https://pith.science/pith/WLAIOSUOK25TRUASQSELYGSOL3/action/author_attestation","sign_citation":"https://pith.science/pith/WLAIOSUOK25TRUASQSELYGSOL3/action/citation_signature","submit_replication":"https://pith.science/pith/WLAIOSUOK25TRUASQSELYGSOL3/action/replication_record"}},"created_at":"2026-05-18T00:05:55.215113+00:00","updated_at":"2026-05-18T00:05:55.215113+00:00"}