{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2022:U6MMRFXZ56C2PV2LGYBCEPW4TT","short_pith_number":"pith:U6MMRFXZ","schema_version":"1.0","canonical_sha256":"a798c896f9ef85a7d74b3602223edc9cfe8aaa4f9e1b0b99f168c16beb732da3","source":{"kind":"arxiv","id":"2205.15213","version":3},"attestation_state":"computed","paper":{"title":"Backpropagation through Combinatorial Algorithms: Identity with Projection Works","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Anselm Paulus, Georg Martius, Marin Vlastelica, Subham Sekhar Sahoo, V\\'it Musil, Volodymyr Kuleshov","submitted_at":"2022-05-30T16:17:09Z","abstract_excerpt":"Embedding discrete solvers as differentiable layers has given modern deep learning architectures combinatorial expressivity and discrete reasoning capabilities. The derivative of these solvers is zero or undefined, therefore a meaningful replacement is crucial for effective gradient-based learning. Prior works rely on smoothing the solver with input perturbations, relaxing the solver to continuous problems, or interpolating the loss landscape with techniques that typically require additional solver calls, introduce extra hyper-parameters, or compromise performance. We propose a principled appr"},"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":"2205.15213","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2022-05-30T16:17:09Z","cross_cats_sorted":[],"title_canon_sha256":"1f292d561db64494799fa64482140fa0d05c7815783ffffc32208ed0313915c5","abstract_canon_sha256":"2a3bc0442d79657278ab8dadd32131be5dac37675e20a4651f3aa5dcf1cc8a67"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:48:53.557551Z","signature_b64":"f51UlVsnTrjbAoTRWU6NZAL4i35uqMuysBVYtQbUE77MUS0x0VEFbA7Y9t0cX0ihtR22VoXAoppqR9SbIIk2Dg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"a798c896f9ef85a7d74b3602223edc9cfe8aaa4f9e1b0b99f168c16beb732da3","last_reissued_at":"2026-07-05T09:48:53.557080Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:48:53.557080Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Backpropagation through Combinatorial Algorithms: Identity with Projection Works","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Anselm Paulus, Georg Martius, Marin Vlastelica, Subham Sekhar Sahoo, V\\'it Musil, Volodymyr Kuleshov","submitted_at":"2022-05-30T16:17:09Z","abstract_excerpt":"Embedding discrete solvers as differentiable layers has given modern deep learning architectures combinatorial expressivity and discrete reasoning capabilities. The derivative of these solvers is zero or undefined, therefore a meaningful replacement is crucial for effective gradient-based learning. Prior works rely on smoothing the solver with input perturbations, relaxing the solver to continuous problems, or interpolating the loss landscape with techniques that typically require additional solver calls, introduce extra hyper-parameters, or compromise performance. We propose a principled appr"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2205.15213","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":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2205.15213/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":"2205.15213","created_at":"2026-07-05T09:48:53.557129+00:00"},{"alias_kind":"arxiv_version","alias_value":"2205.15213v3","created_at":"2026-07-05T09:48:53.557129+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2205.15213","created_at":"2026-07-05T09:48:53.557129+00:00"},{"alias_kind":"pith_short_12","alias_value":"U6MMRFXZ56C2","created_at":"2026-07-05T09:48:53.557129+00:00"},{"alias_kind":"pith_short_16","alias_value":"U6MMRFXZ56C2PV2L","created_at":"2026-07-05T09:48:53.557129+00:00"},{"alias_kind":"pith_short_8","alias_value":"U6MMRFXZ","created_at":"2026-07-05T09:48:53.557129+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":3,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.02294","citing_title":"Regularized Large Neighborhood Search","ref_index":36,"is_internal_anchor":false},{"citing_arxiv_id":"2605.12754","citing_title":"Constraint-Aware Flow Matching: Decision Aligned End-to-End Training for Constrained Sampling","ref_index":43,"is_internal_anchor":false},{"citing_arxiv_id":"2605.01361","citing_title":"Decision-Focused Learning via Tangent-Space Projection of Prediction Error","ref_index":49,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/U6MMRFXZ56C2PV2LGYBCEPW4TT","json":"https://pith.science/pith/U6MMRFXZ56C2PV2LGYBCEPW4TT.json","graph_json":"https://pith.science/api/pith-number/U6MMRFXZ56C2PV2LGYBCEPW4TT/graph.json","events_json":"https://pith.science/api/pith-number/U6MMRFXZ56C2PV2LGYBCEPW4TT/events.json","paper":"https://pith.science/paper/U6MMRFXZ"},"agent_actions":{"view_html":"https://pith.science/pith/U6MMRFXZ56C2PV2LGYBCEPW4TT","download_json":"https://pith.science/pith/U6MMRFXZ56C2PV2LGYBCEPW4TT.json","view_paper":"https://pith.science/paper/U6MMRFXZ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2205.15213&json=true","fetch_graph":"https://pith.science/api/pith-number/U6MMRFXZ56C2PV2LGYBCEPW4TT/graph.json","fetch_events":"https://pith.science/api/pith-number/U6MMRFXZ56C2PV2LGYBCEPW4TT/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/U6MMRFXZ56C2PV2LGYBCEPW4TT/action/timestamp_anchor","attest_storage":"https://pith.science/pith/U6MMRFXZ56C2PV2LGYBCEPW4TT/action/storage_attestation","attest_author":"https://pith.science/pith/U6MMRFXZ56C2PV2LGYBCEPW4TT/action/author_attestation","sign_citation":"https://pith.science/pith/U6MMRFXZ56C2PV2LGYBCEPW4TT/action/citation_signature","submit_replication":"https://pith.science/pith/U6MMRFXZ56C2PV2LGYBCEPW4TT/action/replication_record"}},"created_at":"2026-07-05T09:48:53.557129+00:00","updated_at":"2026-07-05T09:48:53.557129+00:00"}