{"paper":{"title":"Implicit Neural Optimal Transport via Fixed-Point Optimization","license":"http://creativecommons.org/licenses/by/4.0/","headline":"A single neural network solves optimal transport by reformulating the c-transform as a proximal fixed-point problem, enforcing dual feasibility exactly without adversarial training or implicit differentiation.","cross_cats":["cs.LG"],"primary_cat":"math.OC","authors_text":"Eric Gelphman, Samy Wu Fung, Stanley Osher, Yesom Park","submitted_at":"2026-05-11T16:22:06Z","abstract_excerpt":"We propose an implicit neural formulation of optimal transport that eliminates adversarial min--max optimization and multi-network architectures commonly used in existing approaches. Our key idea is to parameterize a single potential in the Kantorovich dual and reformulate the associated c-transform as a proximal fixed-point problem. This yields a stable single-network framework in which dual feasibility is enforced exactly through proximal optimality conditions rather than adversarial training. Despite the inner fixed-point computation, gradients can be computed without differentiating throug"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"We propose an implicit neural formulation of optimal transport that eliminates adversarial min--max optimization and multi-network architectures commonly used in existing approaches. Our key idea is to parameterize a single potential in the Kantorovich dual and reformulate the associated c-transform as a proximal fixed-point problem.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That the proximal fixed-point reformulation of the c-transform can be solved accurately enough in practice to enforce dual feasibility exactly and that gradients computed without implicit differentiation remain faithful to the true optimal transport objective.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"A single-network fixed-point formulation for neural optimal transport eliminates adversarial min-max optimization and implicit differentiation while enforcing dual feasibility exactly.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"A single neural network solves optimal transport by reformulating the c-transform as a proximal fixed-point problem, enforcing dual feasibility exactly without adversarial training or implicit differentiation.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"06c1b1ad1084083c94db86ef60d6a66226dfa37c44c41bdd47fe6438583b8f5f"},"source":{"id":"2605.10792","kind":"arxiv","version":2},"verdict":{"id":"020dc055-777a-4f58-ab63-84ff25c5bc6a","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-12T03:31:10.426088Z","strongest_claim":"We propose an implicit neural formulation of optimal transport that eliminates adversarial min--max optimization and multi-network architectures commonly used in existing approaches. Our key idea is to parameterize a single potential in the Kantorovich dual and reformulate the associated c-transform as a proximal fixed-point problem.","one_line_summary":"A single-network fixed-point formulation for neural optimal transport eliminates adversarial min-max optimization and implicit differentiation while enforcing dual feasibility exactly.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That the proximal fixed-point reformulation of the c-transform can be solved accurately enough in practice to enforce dual feasibility exactly and that gradients computed without implicit differentiation remain faithful to the true optimal transport objective.","pith_extraction_headline":"A single neural network solves optimal transport by reformulating the c-transform as a proximal fixed-point problem, enforcing dual feasibility exactly without adversarial training or implicit differentiation."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2605.10792/integrity.json","findings":[],"available":true,"detectors_run":[{"name":"claim_evidence","ran_at":"2026-05-20T05:22:00.379229Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"ai_meta_artifact","ran_at":"2026-05-19T14:35:35.193495Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"doi_title_agreement","ran_at":"2026-05-19T10:31:17.772124Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"doi_compliance","ran_at":"2026-05-19T08:58:51.781274Z","status":"completed","version":"1.0.0","findings_count":0}],"snapshot_sha256":"5eacecc0d7c7bec1d28a217a1ad610424c8fe82bfd37b920795977770dc949ed"},"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"}