{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:JJQYMCXELAIKJPAAPR4IBVHUWS","short_pith_number":"pith:JJQYMCXE","schema_version":"1.0","canonical_sha256":"4a61860ae45810a4bc007c7880d4f4b4beac65be8f734efefa46f392f6e2bfef","source":{"kind":"arxiv","id":"2305.16368","version":3},"attestation_state":"computed","paper":{"title":"Neural incomplete factorization: learning preconditioners for the conjugate gradient method","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG","cs.NA","math.NA","stat.ML"],"primary_cat":"math.OC","authors_text":"Jens Sj\\\"olund, Ozan \\\"Oktem, Paul H\\\"ausner","submitted_at":"2023-05-25T11:45:46Z","abstract_excerpt":"The convergence of the conjugate gradient method for solving large-scale and sparse linear equation systems depends on the spectral properties of the system matrix, which can be improved by preconditioning. In this paper, we develop a computationally efficient data-driven approach to accelerate the generation of effective preconditioners. We, therefore, replace the typically hand-engineered preconditioners by the output of graph neural networks. Our method generates an incomplete factorization of the matrix and is, therefore, referred to as neural incomplete factorization (NeuralIF). Optimizin"},"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":"2305.16368","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"math.OC","submitted_at":"2023-05-25T11:45:46Z","cross_cats_sorted":["cs.LG","cs.NA","math.NA","stat.ML"],"title_canon_sha256":"d3b1070d88e94f58a5af204e781c68736f9c42a74eae5d3624ad9eaef519272a","abstract_canon_sha256":"77c778acb463e5728c4fb5f5cda02cd31194afb9fa403296fdd4418f052ab12c"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:24:51.065999Z","signature_b64":"VvjaUjWi7hT3uayKPI6MamEB5UBGhlAeUs75eCC9YdFOCckN1+NhiZryorcoeJf9FV+ZPjHxF7PVaNunEF8RAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"4a61860ae45810a4bc007c7880d4f4b4beac65be8f734efefa46f392f6e2bfef","last_reissued_at":"2026-07-05T09:24:51.065505Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:24:51.065505Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Neural incomplete factorization: learning preconditioners for the conjugate gradient method","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG","cs.NA","math.NA","stat.ML"],"primary_cat":"math.OC","authors_text":"Jens Sj\\\"olund, Ozan \\\"Oktem, Paul H\\\"ausner","submitted_at":"2023-05-25T11:45:46Z","abstract_excerpt":"The convergence of the conjugate gradient method for solving large-scale and sparse linear equation systems depends on the spectral properties of the system matrix, which can be improved by preconditioning. In this paper, we develop a computationally efficient data-driven approach to accelerate the generation of effective preconditioners. We, therefore, replace the typically hand-engineered preconditioners by the output of graph neural networks. Our method generates an incomplete factorization of the matrix and is, therefore, referred to as neural incomplete factorization (NeuralIF). Optimizin"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2305.16368","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/2305.16368/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":"2305.16368","created_at":"2026-07-05T09:24:51.065562+00:00"},{"alias_kind":"arxiv_version","alias_value":"2305.16368v3","created_at":"2026-07-05T09:24:51.065562+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2305.16368","created_at":"2026-07-05T09:24:51.065562+00:00"},{"alias_kind":"pith_short_12","alias_value":"JJQYMCXELAIK","created_at":"2026-07-05T09:24:51.065562+00:00"},{"alias_kind":"pith_short_16","alias_value":"JJQYMCXELAIKJPAA","created_at":"2026-07-05T09:24:51.065562+00:00"},{"alias_kind":"pith_short_8","alias_value":"JJQYMCXE","created_at":"2026-07-05T09:24:51.065562+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":3,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.22742","citing_title":"Factored Sparse Approximate Inverse Preconditioning via Spectral Optimization","ref_index":74,"is_internal_anchor":false},{"citing_arxiv_id":"2605.18004","citing_title":"RL4RLA: Teaching ML to Discover Randomized Linear Algebra Algorithms Through Curriculum Design and Graph-Based Search","ref_index":66,"is_internal_anchor":false},{"citing_arxiv_id":"2605.19867","citing_title":"When can a neural operator replace a coarse solve? Architectural principles for two-level preconditioning","ref_index":20,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/JJQYMCXELAIKJPAAPR4IBVHUWS","json":"https://pith.science/pith/JJQYMCXELAIKJPAAPR4IBVHUWS.json","graph_json":"https://pith.science/api/pith-number/JJQYMCXELAIKJPAAPR4IBVHUWS/graph.json","events_json":"https://pith.science/api/pith-number/JJQYMCXELAIKJPAAPR4IBVHUWS/events.json","paper":"https://pith.science/paper/JJQYMCXE"},"agent_actions":{"view_html":"https://pith.science/pith/JJQYMCXELAIKJPAAPR4IBVHUWS","download_json":"https://pith.science/pith/JJQYMCXELAIKJPAAPR4IBVHUWS.json","view_paper":"https://pith.science/paper/JJQYMCXE","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2305.16368&json=true","fetch_graph":"https://pith.science/api/pith-number/JJQYMCXELAIKJPAAPR4IBVHUWS/graph.json","fetch_events":"https://pith.science/api/pith-number/JJQYMCXELAIKJPAAPR4IBVHUWS/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/JJQYMCXELAIKJPAAPR4IBVHUWS/action/timestamp_anchor","attest_storage":"https://pith.science/pith/JJQYMCXELAIKJPAAPR4IBVHUWS/action/storage_attestation","attest_author":"https://pith.science/pith/JJQYMCXELAIKJPAAPR4IBVHUWS/action/author_attestation","sign_citation":"https://pith.science/pith/JJQYMCXELAIKJPAAPR4IBVHUWS/action/citation_signature","submit_replication":"https://pith.science/pith/JJQYMCXELAIKJPAAPR4IBVHUWS/action/replication_record"}},"created_at":"2026-07-05T09:24:51.065562+00:00","updated_at":"2026-07-05T09:24:51.065562+00:00"}