{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:RQY4AMNL4FKBQJ5K3MDEHL4HVB","short_pith_number":"pith:RQY4AMNL","schema_version":"1.0","canonical_sha256":"8c31c031abe1541827aadb0643af87a872e90cabce9323cdc749895c610c3956","source":{"kind":"arxiv","id":"2603.10395","version":2},"attestation_state":"computed","paper":{"title":"Graph-GRPO: Training Graph Flow Models with Reinforcement Learning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Baoheng Zhu, Delvin Ce Zhang, Deyu Bo, Xiao Wang","submitted_at":"2026-03-11T04:20:45Z","abstract_excerpt":"Graph generation is a fundamental task with broad applications, such as drug discovery. Recently, discrete flow matching-based graph generation, \\aka, graph flow model (GFM), has emerged due to its superior performance and flexible sampling. However, effectively aligning GFMs with complex human preferences or task-specific objectives remains a significant challenge. In this paper, we propose Graph-GRPO, an online reinforcement learning (RL) framework for training GFMs under verifiable rewards. Our method makes two key contributions: (1) We derive an analytical expression for the transition pro"},"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":"2603.10395","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-03-11T04:20:45Z","cross_cats_sorted":[],"title_canon_sha256":"f342c3c3f1f390397aec318a27654ba6b667368d9d1aefb381190088a788aa89","abstract_canon_sha256":"8c5814c8d7d1a827557008382bcfd5260a4717f064b65cc4ffb9b750df4e9df4"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-09T02:07:23.587358Z","signature_b64":"Pmwt+E/k6Vs0jOPSApsFi1Cy7e+I3V6FGof2MCPGETsAkfwam54AUrlE5edop445PMAmcxmjUZqcK2z5+iJPCQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"8c31c031abe1541827aadb0643af87a872e90cabce9323cdc749895c610c3956","last_reissued_at":"2026-06-09T02:07:23.586365Z","signature_status":"signed_v1","first_computed_at":"2026-06-09T02:07:23.586365Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Graph-GRPO: Training Graph Flow Models with Reinforcement Learning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Baoheng Zhu, Delvin Ce Zhang, Deyu Bo, Xiao Wang","submitted_at":"2026-03-11T04:20:45Z","abstract_excerpt":"Graph generation is a fundamental task with broad applications, such as drug discovery. Recently, discrete flow matching-based graph generation, \\aka, graph flow model (GFM), has emerged due to its superior performance and flexible sampling. However, effectively aligning GFMs with complex human preferences or task-specific objectives remains a significant challenge. In this paper, we propose Graph-GRPO, an online reinforcement learning (RL) framework for training GFMs under verifiable rewards. Our method makes two key contributions: (1) We derive an analytical expression for the transition pro"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2603.10395","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/2603.10395/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":"2603.10395","created_at":"2026-06-09T02:07:23.586516+00:00"},{"alias_kind":"arxiv_version","alias_value":"2603.10395v2","created_at":"2026-06-09T02:07:23.586516+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2603.10395","created_at":"2026-06-09T02:07:23.586516+00:00"},{"alias_kind":"pith_short_12","alias_value":"RQY4AMNL4FKB","created_at":"2026-06-09T02:07:23.586516+00:00"},{"alias_kind":"pith_short_16","alias_value":"RQY4AMNL4FKBQJ5K","created_at":"2026-06-09T02:07:23.586516+00:00"},{"alias_kind":"pith_short_8","alias_value":"RQY4AMNL","created_at":"2026-06-09T02:07:23.586516+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2605.14483","citing_title":"LEMON: Learning Executable Multi-Agent Orchestration via Counterfactual Reinforcement Learning","ref_index":25,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/RQY4AMNL4FKBQJ5K3MDEHL4HVB","json":"https://pith.science/pith/RQY4AMNL4FKBQJ5K3MDEHL4HVB.json","graph_json":"https://pith.science/api/pith-number/RQY4AMNL4FKBQJ5K3MDEHL4HVB/graph.json","events_json":"https://pith.science/api/pith-number/RQY4AMNL4FKBQJ5K3MDEHL4HVB/events.json","paper":"https://pith.science/paper/RQY4AMNL"},"agent_actions":{"view_html":"https://pith.science/pith/RQY4AMNL4FKBQJ5K3MDEHL4HVB","download_json":"https://pith.science/pith/RQY4AMNL4FKBQJ5K3MDEHL4HVB.json","view_paper":"https://pith.science/paper/RQY4AMNL","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2603.10395&json=true","fetch_graph":"https://pith.science/api/pith-number/RQY4AMNL4FKBQJ5K3MDEHL4HVB/graph.json","fetch_events":"https://pith.science/api/pith-number/RQY4AMNL4FKBQJ5K3MDEHL4HVB/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/RQY4AMNL4FKBQJ5K3MDEHL4HVB/action/timestamp_anchor","attest_storage":"https://pith.science/pith/RQY4AMNL4FKBQJ5K3MDEHL4HVB/action/storage_attestation","attest_author":"https://pith.science/pith/RQY4AMNL4FKBQJ5K3MDEHL4HVB/action/author_attestation","sign_citation":"https://pith.science/pith/RQY4AMNL4FKBQJ5K3MDEHL4HVB/action/citation_signature","submit_replication":"https://pith.science/pith/RQY4AMNL4FKBQJ5K3MDEHL4HVB/action/replication_record"}},"created_at":"2026-06-09T02:07:23.586516+00:00","updated_at":"2026-06-09T02:07:23.586516+00:00"}