{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2017:EHNJZ4RJENP4JXVPWF7ACUIEDR","short_pith_number":"pith:EHNJZ4RJ","schema_version":"1.0","canonical_sha256":"21da9cf229235fc4deafb17e0151041c5a1b7e3af953911c758f76596a04baa5","source":{"kind":"arxiv","id":"1706.02025","version":1},"attestation_state":"computed","paper":{"title":"Imposing Hard Constraints on Deep Networks: Promises and Limitations","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Mathieu Salzmann, Pablo M\\'arquez-Neila, Pascal Fua","submitted_at":"2017-06-07T02:03:00Z","abstract_excerpt":"Imposing constraints on the output of a Deep Neural Net is one way to improve the quality of its predictions while loosening the requirements for labeled training data. Such constraints are usually imposed as soft constraints by adding new terms to the loss function that is minimized during training. An alternative is to impose them as hard constraints, which has a number of theoretical benefits but has not been explored so far due to the perceived intractability of the problem.\n  In this paper, we show that imposing hard constraints can in fact be done in a computationally feasible way and de"},"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":"1706.02025","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2017-06-07T02:03:00Z","cross_cats_sorted":[],"title_canon_sha256":"dc7b9e6505ad811676eefe47f866f16a0e58831217a4b9fcea1ff1044f04ab6d","abstract_canon_sha256":"f68e0d9baea8ed0193840dc26db0d3241b51e96c6c30192fed2e92a230df6c5c"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-18T00:42:50.443299Z","signature_b64":"SV5UFW68Ez3pSKdFXE94kPnUukNgG9QnqSgK0QaOW35HjrqD2uH3ZtIoHFLWEZ/9DyeXFx7d0oCQ5r9duNoQAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"21da9cf229235fc4deafb17e0151041c5a1b7e3af953911c758f76596a04baa5","last_reissued_at":"2026-05-18T00:42:50.442711Z","signature_status":"signed_v1","first_computed_at":"2026-05-18T00:42:50.442711Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Imposing Hard Constraints on Deep Networks: Promises and Limitations","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Mathieu Salzmann, Pablo M\\'arquez-Neila, Pascal Fua","submitted_at":"2017-06-07T02:03:00Z","abstract_excerpt":"Imposing constraints on the output of a Deep Neural Net is one way to improve the quality of its predictions while loosening the requirements for labeled training data. Such constraints are usually imposed as soft constraints by adding new terms to the loss function that is minimized during training. An alternative is to impose them as hard constraints, which has a number of theoretical benefits but has not been explored so far due to the perceived intractability of the problem.\n  In this paper, we show that imposing hard constraints can in fact be done in a computationally feasible way and de"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1706.02025","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":"1706.02025","created_at":"2026-05-18T00:42:50.442832+00:00"},{"alias_kind":"arxiv_version","alias_value":"1706.02025v1","created_at":"2026-05-18T00:42:50.442832+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1706.02025","created_at":"2026-05-18T00:42:50.442832+00:00"},{"alias_kind":"pith_short_12","alias_value":"EHNJZ4RJENP4","created_at":"2026-05-18T12:31:12.930513+00:00"},{"alias_kind":"pith_short_16","alias_value":"EHNJZ4RJENP4JXVP","created_at":"2026-05-18T12:31:12.930513+00:00"},{"alias_kind":"pith_short_8","alias_value":"EHNJZ4RJ","created_at":"2026-05-18T12:31:12.930513+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":4,"internal_anchor_count":3,"sample":[{"citing_arxiv_id":"2307.08336","citing_title":"RAYEN: Imposition of Hard Convex Constraints on Neural Networks","ref_index":7,"is_internal_anchor":true},{"citing_arxiv_id":"2605.16383","citing_title":"A neurosymbolic Approach with Epistemic Deep Learning for Hierarchical Image Classification","ref_index":100,"is_internal_anchor":true},{"citing_arxiv_id":"2602.09317","citing_title":"SnareNet: Flexible Repair Layers for Neural Networks with Hard Constraints","ref_index":29,"is_internal_anchor":true},{"citing_arxiv_id":"2605.00260","citing_title":"NLPOpt-Net: A Learning Method for Nonlinear Optimization with Feasibility Guarantees","ref_index":12,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/EHNJZ4RJENP4JXVPWF7ACUIEDR","json":"https://pith.science/pith/EHNJZ4RJENP4JXVPWF7ACUIEDR.json","graph_json":"https://pith.science/api/pith-number/EHNJZ4RJENP4JXVPWF7ACUIEDR/graph.json","events_json":"https://pith.science/api/pith-number/EHNJZ4RJENP4JXVPWF7ACUIEDR/events.json","paper":"https://pith.science/paper/EHNJZ4RJ"},"agent_actions":{"view_html":"https://pith.science/pith/EHNJZ4RJENP4JXVPWF7ACUIEDR","download_json":"https://pith.science/pith/EHNJZ4RJENP4JXVPWF7ACUIEDR.json","view_paper":"https://pith.science/paper/EHNJZ4RJ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=1706.02025&json=true","fetch_graph":"https://pith.science/api/pith-number/EHNJZ4RJENP4JXVPWF7ACUIEDR/graph.json","fetch_events":"https://pith.science/api/pith-number/EHNJZ4RJENP4JXVPWF7ACUIEDR/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/EHNJZ4RJENP4JXVPWF7ACUIEDR/action/timestamp_anchor","attest_storage":"https://pith.science/pith/EHNJZ4RJENP4JXVPWF7ACUIEDR/action/storage_attestation","attest_author":"https://pith.science/pith/EHNJZ4RJENP4JXVPWF7ACUIEDR/action/author_attestation","sign_citation":"https://pith.science/pith/EHNJZ4RJENP4JXVPWF7ACUIEDR/action/citation_signature","submit_replication":"https://pith.science/pith/EHNJZ4RJENP4JXVPWF7ACUIEDR/action/replication_record"}},"created_at":"2026-05-18T00:42:50.442832+00:00","updated_at":"2026-05-18T00:42:50.442832+00:00"}