{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:BZCPENM5EK6U35XRGUGAGNVYEW","short_pith_number":"pith:BZCPENM5","schema_version":"1.0","canonical_sha256":"0e44f2359d22bd4df6f1350c0336b825b0ba24c4e9ad88a9473ef475f0e190f2","source":{"kind":"arxiv","id":"2403.15180","version":2},"attestation_state":"computed","paper":{"title":"Self-Improvement for Neural Combinatorial Optimization: Sample without Replacement, but Improvement","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Dominik G. Grimm, Jonathan Pirnay","submitted_at":"2024-03-22T13:09:10Z","abstract_excerpt":"Current methods for end-to-end constructive neural combinatorial optimization usually train a policy using behavior cloning from expert solutions or policy gradient methods from reinforcement learning. While behavior cloning is straightforward, it requires expensive expert solutions, and policy gradient methods are often computationally demanding and complex to fine-tune. In this work, we bridge the two and simplify the training process by sampling multiple solutions for random instances using the current model in each epoch and then selecting the best solution as an expert trajectory for supe"},"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":"2403.15180","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-03-22T13:09:10Z","cross_cats_sorted":[],"title_canon_sha256":"93dee9a2a289c0fbfc29562a0a03a6680adac03ddc8fae155d8a2970994c986b","abstract_canon_sha256":"792cc92ac4ef284b61ea66c7515c32e765c95af74c0a1588b81246906e1b0996"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:30:11.234360Z","signature_b64":"SfJwTtkC1JKkyfSuhAc8mDaaGpjFjGzo/Vaqm2pvRXmuLEdDuDmvPNJyCLTahiqzArxCd3wQc930fV6BNf8MAQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"0e44f2359d22bd4df6f1350c0336b825b0ba24c4e9ad88a9473ef475f0e190f2","last_reissued_at":"2026-07-05T09:30:11.233669Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:30:11.233669Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Self-Improvement for Neural Combinatorial Optimization: Sample without Replacement, but Improvement","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Dominik G. Grimm, Jonathan Pirnay","submitted_at":"2024-03-22T13:09:10Z","abstract_excerpt":"Current methods for end-to-end constructive neural combinatorial optimization usually train a policy using behavior cloning from expert solutions or policy gradient methods from reinforcement learning. While behavior cloning is straightforward, it requires expensive expert solutions, and policy gradient methods are often computationally demanding and complex to fine-tune. In this work, we bridge the two and simplify the training process by sampling multiple solutions for random instances using the current model in each epoch and then selecting the best solution as an expert trajectory for supe"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2403.15180","kind":"arxiv","version":2},"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/2403.15180/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":"2403.15180","created_at":"2026-07-05T09:30:11.233798+00:00"},{"alias_kind":"arxiv_version","alias_value":"2403.15180v2","created_at":"2026-07-05T09:30:11.233798+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2403.15180","created_at":"2026-07-05T09:30:11.233798+00:00"},{"alias_kind":"pith_short_12","alias_value":"BZCPENM5EK6U","created_at":"2026-07-05T09:30:11.233798+00:00"},{"alias_kind":"pith_short_16","alias_value":"BZCPENM5EK6U35XR","created_at":"2026-07-05T09:30:11.233798+00:00"},{"alias_kind":"pith_short_8","alias_value":"BZCPENM5","created_at":"2026-07-05T09:30:11.233798+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.26690","citing_title":"Self-Improvement Imitation with Biologically Guided Search for Protein Design Under Oracle Budgets","ref_index":32,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/BZCPENM5EK6U35XRGUGAGNVYEW","json":"https://pith.science/pith/BZCPENM5EK6U35XRGUGAGNVYEW.json","graph_json":"https://pith.science/api/pith-number/BZCPENM5EK6U35XRGUGAGNVYEW/graph.json","events_json":"https://pith.science/api/pith-number/BZCPENM5EK6U35XRGUGAGNVYEW/events.json","paper":"https://pith.science/paper/BZCPENM5"},"agent_actions":{"view_html":"https://pith.science/pith/BZCPENM5EK6U35XRGUGAGNVYEW","download_json":"https://pith.science/pith/BZCPENM5EK6U35XRGUGAGNVYEW.json","view_paper":"https://pith.science/paper/BZCPENM5","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2403.15180&json=true","fetch_graph":"https://pith.science/api/pith-number/BZCPENM5EK6U35XRGUGAGNVYEW/graph.json","fetch_events":"https://pith.science/api/pith-number/BZCPENM5EK6U35XRGUGAGNVYEW/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/BZCPENM5EK6U35XRGUGAGNVYEW/action/timestamp_anchor","attest_storage":"https://pith.science/pith/BZCPENM5EK6U35XRGUGAGNVYEW/action/storage_attestation","attest_author":"https://pith.science/pith/BZCPENM5EK6U35XRGUGAGNVYEW/action/author_attestation","sign_citation":"https://pith.science/pith/BZCPENM5EK6U35XRGUGAGNVYEW/action/citation_signature","submit_replication":"https://pith.science/pith/BZCPENM5EK6U35XRGUGAGNVYEW/action/replication_record"}},"created_at":"2026-07-05T09:30:11.233798+00:00","updated_at":"2026-07-05T09:30:11.233798+00:00"}