{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:63Q6LYCDVJHTVNWPCVZBI3QV4L","short_pith_number":"pith:63Q6LYCD","schema_version":"1.0","canonical_sha256":"f6e1e5e043aa4f3ab6cf1572146e15e2f21d96ce9d94b21fdb3d03918bcc406d","source":{"kind":"arxiv","id":"2605.13673","version":1},"attestation_state":"computed","paper":{"title":"Graph Neural Networks with Triangle-Based Messages for the Multicut Problem","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"A graph neural network that passes messages only along triangles in edge-featured graphs finds higher-quality solutions to the multicut problem than existing heuristic solvers.","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Bjoern Andres, Jannik Irmai, Lucas Fabian Naumann","submitted_at":"2026-05-13T15:33:13Z","abstract_excerpt":"The multicut problem is an NP-hard combinatorial optimization problem with diverse applications in fields such as bioinformatics, data mining and computer vision. Graph neural networks have been defined for the multicut problem but can be adapted further to its specific objective function and constraints. In this article, we introduce such an adapted graph neural network architecture in which features are assigned only to edges, and the computation of messages is based on triangles in the underlying graph. Experiments with synthetic and real-world instances with up to 200 nodes show that our m"},"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":true,"formal_links_present":false},"canonical_record":{"source":{"id":"2605.13673","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-05-13T15:33:13Z","cross_cats_sorted":[],"title_canon_sha256":"69ac531c67a3e5f79021cb24365125878aa8bfdaf19a8b66adef4501cd250282","abstract_canon_sha256":"bc17fd0e50da1cae88468506639d1a7cc2a2ced156d7406118cfaeb71d16f07b"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-18T02:44:17.142536Z","signature_b64":"VPXPlMsuxVwy33nxgWFEZ0pT/NBIydIeq25UB8lzgsl/HmN3MlSq4J8NlaUF/5PGZfOHNujUAkRec+H01G0yAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"f6e1e5e043aa4f3ab6cf1572146e15e2f21d96ce9d94b21fdb3d03918bcc406d","last_reissued_at":"2026-05-18T02:44:17.142046Z","signature_status":"signed_v1","first_computed_at":"2026-05-18T02:44:17.142046Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Graph Neural Networks with Triangle-Based Messages for the Multicut Problem","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"A graph neural network that passes messages only along triangles in edge-featured graphs finds higher-quality solutions to the multicut problem than existing heuristic solvers.","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Bjoern Andres, Jannik Irmai, Lucas Fabian Naumann","submitted_at":"2026-05-13T15:33:13Z","abstract_excerpt":"The multicut problem is an NP-hard combinatorial optimization problem with diverse applications in fields such as bioinformatics, data mining and computer vision. Graph neural networks have been defined for the multicut problem but can be adapted further to its specific objective function and constraints. In this article, we introduce such an adapted graph neural network architecture in which features are assigned only to edges, and the computation of messages is based on triangles in the underlying graph. Experiments with synthetic and real-world instances with up to 200 nodes show that our m"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"Experiments with synthetic and real-world instances with up to 200 nodes show that our method outperforms state-of-the-art heuristic solvers in terms of solution quality while maintaining feasible runtimes. For some instances, our method finds optimal solutions in seconds whereas exact solvers need hours to find and certify optimal solutions.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"The assumption that training the triangle-message GNN on the provided synthetic and real-world instances produces a model that generalizes reliably to new instances of the multicut problem without overfitting to the training distribution or requiring extensive hyperparameter tuning.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"A triangle-message GNN for multicut outperforms heuristics in solution quality on graphs up to 200 nodes and finds optimal solutions faster than exact solvers for some cases.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"A graph neural network that passes messages only along triangles in edge-featured graphs finds higher-quality solutions to the multicut problem than existing heuristic solvers.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"cad7547c689d113a06e94ff4d480139830e9d59acc345db5aa0feb3720c61b56"},"source":{"id":"2605.13673","kind":"arxiv","version":1},"verdict":{"id":"3060c708-2339-4abb-a913-e1ccff82f746","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-14T19:12:17.291595Z","strongest_claim":"Experiments with synthetic and real-world instances with up to 200 nodes show that our method outperforms state-of-the-art heuristic solvers in terms of solution quality while maintaining feasible runtimes. For some instances, our method finds optimal solutions in seconds whereas exact solvers need hours to find and certify optimal solutions.","one_line_summary":"A triangle-message GNN for multicut outperforms heuristics in solution quality on graphs up to 200 nodes and finds optimal solutions faster than exact solvers for some cases.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"The assumption that training the triangle-message GNN on the provided synthetic and real-world instances produces a model that generalizes reliably to new instances of the multicut problem without overfitting to the training distribution or requiring extensive hyperparameter tuning.","pith_extraction_headline":"A graph neural network that passes messages only along triangles in edge-featured graphs finds higher-quality solutions to the multicut problem than existing heuristic solvers."},"references":{"count":42,"sample":[{"doi":"10.1007/978-3-031-26390-3_28","year":2023,"title":"Learning to Solve Minimum Cost Multicuts Efficiently Using Edge-Weighted Graph Convolutional Neural Networks","work_id":"ced593bd-22b6-4811-9335-fb5196f7e8ac","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2017,"title":"and Riley, Patrick F","work_id":"145cc425-b730-481f-93c4-f97a57ff111f","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2022,"title":"2022 , url =","work_id":"cad3f236-417a-4cf1-a97c-963639c7b4f6","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2025,"title":"A State-of-the-Art Cutting Plane Algorithm for Clique Partitioning","work_id":"872fb60d-c2b5-4852-a00c-d90c2da69cee","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":null,"title":"Mathematical Programming Computation , year=","work_id":"e07a3d03-86aa-4d94-adff-ceab8288912c","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":42,"snapshot_sha256":"2c1a08bd6dc7857a5c43996f69f7fd3349c68b27974ac506ee98f40474f28511","internal_anchors":1},"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":"2605.13673","created_at":"2026-05-18T02:44:17.142130+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.13673v1","created_at":"2026-05-18T02:44:17.142130+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.13673","created_at":"2026-05-18T02:44:17.142130+00:00"},{"alias_kind":"pith_short_12","alias_value":"63Q6LYCDVJHT","created_at":"2026-05-18T12:33:37.589309+00:00"},{"alias_kind":"pith_short_16","alias_value":"63Q6LYCDVJHTVNWP","created_at":"2026-05-18T12:33:37.589309+00:00"},{"alias_kind":"pith_short_8","alias_value":"63Q6LYCD","created_at":"2026-05-18T12:33:37.589309+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2606.11946","citing_title":"Neuro-Relational Programs: Unifying Queries and Neural Computation over Structured Data","ref_index":33,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/63Q6LYCDVJHTVNWPCVZBI3QV4L","json":"https://pith.science/pith/63Q6LYCDVJHTVNWPCVZBI3QV4L.json","graph_json":"https://pith.science/api/pith-number/63Q6LYCDVJHTVNWPCVZBI3QV4L/graph.json","events_json":"https://pith.science/api/pith-number/63Q6LYCDVJHTVNWPCVZBI3QV4L/events.json","paper":"https://pith.science/paper/63Q6LYCD"},"agent_actions":{"view_html":"https://pith.science/pith/63Q6LYCDVJHTVNWPCVZBI3QV4L","download_json":"https://pith.science/pith/63Q6LYCDVJHTVNWPCVZBI3QV4L.json","view_paper":"https://pith.science/paper/63Q6LYCD","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.13673&json=true","fetch_graph":"https://pith.science/api/pith-number/63Q6LYCDVJHTVNWPCVZBI3QV4L/graph.json","fetch_events":"https://pith.science/api/pith-number/63Q6LYCDVJHTVNWPCVZBI3QV4L/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/63Q6LYCDVJHTVNWPCVZBI3QV4L/action/timestamp_anchor","attest_storage":"https://pith.science/pith/63Q6LYCDVJHTVNWPCVZBI3QV4L/action/storage_attestation","attest_author":"https://pith.science/pith/63Q6LYCDVJHTVNWPCVZBI3QV4L/action/author_attestation","sign_citation":"https://pith.science/pith/63Q6LYCDVJHTVNWPCVZBI3QV4L/action/citation_signature","submit_replication":"https://pith.science/pith/63Q6LYCDVJHTVNWPCVZBI3QV4L/action/replication_record"}},"created_at":"2026-05-18T02:44:17.142130+00:00","updated_at":"2026-05-18T02:44:17.142130+00:00"}